Plasmonic Copper: Ways and Means of Achieving, Directing, and Utilizing Surface Plasmons

Abstract

Copper has been reclusive with regard to its plasmonic investigations among the founding plasmonic metals. With the advent of technology and the associated improvements in understanding of plasmonics, copper has been able to make a stand for itself among its peers and even outshine them in a few aspects such as dielectric loss, cost, and a more intricate and facile tuning of the near and far field intensities of the plasmon-enhanced distributions. This review is aimed at highlighting the different classes of plasmonic copper (PC), ranging from its pristine version to the array of composited and alloyed compositions. The focus is on an all-encompassing review of PC with regard to its shortcomings and merits, its exploration for plasmonic applications, and emerging phenomena discovered due to the plasmonic virtue. We aim to bring about a comprehensive treatise of the investigations on PC, where the major discussions are on the topics of a generic treatise on surface plasmons (both localized and propagating), pristine copper and its potential for different applications, the almost inescapable phenomenon of oxidation, and the associations that copper has been made to form in order to be exploited for multiple uses such as chalcogenides, silicides, alloys, and other metamaterial architectures. Specific outcomes of the changes to the near and far-field distributions of PC in various conditions such as oxidized/alloyed/composited and stabilized have been discussed, highlighting the changes to PC in lieu of these modifications. The concluding sections highlight some fascinating compositions including multi-elemental copper and its atomic clusters and cursorily studied compositions which are among the few materials that could offer untapped capabilities which will be made evident from brief glimpses of their plasmonic character. The outlook for plasmonic copper has never been more promising, ranging from the need for comprehensive investigations of emerging material compositions and configurations (of both pristine and composited copper) to the realm of commercialization. Copper has, thus, been projected to be a viable alternative to existing options including the poster children of plasmonics, namely, silver and gold.

Graphical Abstract

Introduction

Plasmonic nanoparticles have wide applications due to their fundamental virtue of being light manipulators, such as in sensing, catalysis, switches, waveguides, heaters, and theranostic agents [1]. Among the coinage metals, Cu has often had to take a back seat with regard to its promise in plasmonic applications due to the superior properties of the other two, among which stability and low optical loss are often attributed to be the reasons for their being frontrunners for plasmon-enhanced properties. As an example of the potentially inferior nature of Cu compared to Au and Ag, the degree of overlap between the inter-(excitonic) and intra-(plasmonic) bands of Cu nanoparticles (NPs) (≈ 0.03 eV) is much higher than Au and Ag, causing the LSPR absorption to be fundamentally less intense. This is all the more evident while considering the energy spacing between the 3 days and 4 s orbitals for Cu in comparison with Au and Ag [2]. Yet, plasmonic copper has rekindled the interest of researchers recently due to its promise as a replacement for Au and Ag in a myriad of applications such as for enhanced scattering, catalysis, and even non-linear optical (NLO) properties. In this review, we discuss how this has become possible with the advent of nanotechnology and nanomaterials. Specifically, with the evolution of synthesis techniques and fabrication technologies, delicate control of the oxidation of Cu is one such enabler, allowing the blooming of investigations aimed at replacing expensive plasmonic metals such as Au and Ag with Cu. We have attempted to highlight the prospect of Cu as a plasmonic metal by discussing its applicability towards every aspect of plasmonic research to the best of our reach of published literature.

We start with a discussion on plasmonics in general and the factors to be mindful of while trying to manipulate plasmonic behavior, with a restriction that only surface plasmon resonance (LSPR) is dealt with and not its companion, bulk plasmon resonance. We then discuss the two themes of research of pristine Cu plasmonics, viz., plasmonic Cu nanoparticles (CuNPs) and films. Attempts to be comprehensive in terms of the prospects of both these configurations have been made by a review of their promise for a wide array of applications. The very interesting and quintessential part of Cu plasmonics, that is, oxidation, is discussed subsequently. Oxidation can be avoided and completely reversed, as multiple investigations have revealed, and can even be beneficial when control is achieved of the oxidation rate and if permanent termination of the oxidation process can be intentionally induced. We then discuss the extensive avenue of Cu composites such as chalcogenides, carbides, silicides, alloys, multi-elemental composites, and other hitherto unexplored (from a plasmonic angle) compositions of Cu. Comprehensive discussions on often explored plasmon-enhanced properties, such as the use of this enhancement for theranostics, catalysis, energy storage, and Raman scattering along with unexplored/rarely explored aspects such as multi-elemental compositions (composed of Cu clusters with more than three elements) and applications such as NLO and electron emitting plasmonic sources, are attempted. As will be demonstrated from this review, the property of Cu being less expensive (and hence amenable for unconventional explorations in terms of composition and for mass production) and having a dielectric function that can be easily (due to inherently high reactivity) and intricately varied for manipulations of the plasmonic profile can make it the metal of choice for plasmonic applications. We conclude this review with a discussion of the myriad of aspects that are unaddressed/those that need a more rigorous confirmation and the avenues that are yet untouched by plasmonic Cu nanostructures, highlighting the conclusion that Cu plasmonics is an increasingly expanding area of the broader field of plasmonics.

SPP vs LSPR: A Brief Treatise

The collective excitation of free electrons at a metal–dielectric interface in response to an electromagnetic field is termed as plasmon resonance. The excitation being dependent intricately on the electronic structure of the material, the material constants such as the dielectric constant, rate constant of different types of excitations that can induce a plasmonic resonance etc. are important, and excellent reviews can be found elucidating the same [3]. This resonant oscillation can be very different when induced in continuous films compared to induction in discrete nanoparticles. In films, this excitation can lead to the generation of a propagating wave of plasmons (the surface plasmon polariton wave or SPP), whereas due to the localization of the excitation within the boundaries of a nanoparticle, it is termed as localized surface plasmon resonance (LSPR) when induced in them. We give here an overall understanding of the theoretical calculations which can help predict the defining attributes of the plasmon resonance such as the damping, the frequency, propagation constant, and the dielectric constants.

For a propagating wave at the interface between a dielectric and the metal, we can consider the situation illustrated in Fig. 1 [4], where 1 denotes the dielectric and 2 is the metal.

Fig. 1

Propagating SPP wave at the interface between a metal and a dielectric

From classical electrodynamics defining the conditions for continuity of a wave between 1 and 2 and from Maxwell’s equations which when solved allow computation of the fields created by the oscillating plasma of electrons, we get the relation between the wave vectors of the SPP in the metal and the dielectric as [4]

$$\frac{{k}_{1,Z}}{{\varepsilon }_{1}}= \frac{{k}_{2,Z}}{{\varepsilon }_{2}}$$

(1)

where k₁, k₂ are wave vectors and ε₁, ε₂ are the dielectric constants.

From this relation, it is evident that the plasmon wave can exist at the interface only for the condition of ε₁ = − ε₂, that the wave exists only at frequencies of negative values of the dielectric function, and that the wave can exist only at the interface on the assumption of an exponential decay away from it, as shown in Fig. 2 [4].

Fig. 2

The field of the propagating plasmon wave at the interface where it decays with different degrees into the respective media

The propagation constant of the SPP is given as [5]

$${k}_{sp}=\frac{\omega }{c}{[\frac{{\varepsilon }_{r}\varepsilon (\omega )}{{\varepsilon }_{r}+\varepsilon (\omega )}]}^{1/2}$$

(2)

where k_sp is the wavenumber of surface plasmon polariton, ε(ω) is the dielectric constant as a function of frequency, and c is the speed of light.

As briefly mentioned, the condition for the SPP constrained to the interface dictates an exponential decay away from it. The penetration depth of the evanescent field of the SPP away from the interface is hence different for the dielectric and the metal, where this distance is referred to as the skin depth for the latter. The value of this propagation length (L_sp) is arrived at from the imaginary part of the SPP wavenumber [5]

$$\frac{1}{{L}_{sp}}=2Im{k}_{sp}=\frac{\omega }{c}\frac{{\varepsilon }_{1}^\frac{3}{2}{\varepsilon }_{2}(\omega )}{{\left|{\varepsilon }_{1}\left(\omega \right)\right|}^\frac{1}{2}({\left|{\varepsilon }_{1}\left(\omega \right)\right|-{\varepsilon }_{1})}^{3/2}}$$

(3)

The discussion so far defining the nature of the SPP once induced, it is important that the SPP also needs to couple with the incoming electromagnetic wave for resonant excitation and hence requires both frequency and wave vector matching with it (in contrast to localized surface plasmons (LSPs) which do not as they are not a propagating wave). This can happen only when the condition of momentum matching is satisfied between the two waves. It is to be kept in mind that the momentum of the SPP is always greater than that of a free space photon, as shown in Fig. 3 [4] which illustrates the dispersion relation of the plasmon wave. Various means of achieving this coupling exist and are broadly classified as through photon tunneling (through the configurations pioneered by Otto and Kretschmann) and through diffraction coupling (through the employment of scattering features such as gratings on the metallic films), although other unconventional methods such as through electron tunneling and LSP scattering have also been reported.

Fig. 3

Dispersion relation between wave vector and frequency of light in a vacuum, in a dielectric, and of surface plasmons

With the defining attributes of the SPP being described until this juncture, we will attempt to describe similar attributes for the second type of resonating plasmons, i.e., LSPs. LSPs are electrostatic oscillations in contrast to propagating SPPs and can arise both in films (on confined areas such as roughness features) and in bound geometries such as nanoparticles. The location and width of LSPs are defined by the particle size/shape, composition, and the dielectric environment around it. In deriving the characteristic oscillation frequency of the LSPs, one starts with calculating the displacement of the collective plasmon oscillation (termed as dielectric displacement, D) in relation to the incident electric field (E) of the incident electromagnetic wave as [6]

$$D = {\varepsilon }_{0}E+P$$

(4)

where P is the polarization density which is

$$P=nex$$

(5)

The number density being (n) and e is the charge of the electron, and the value of x is arrived at by solving the equation of motion of a free electron with mass m, given by [4]

$$m\ddot{x}+m\gamma \dot{x}= -eE$$

(6)

Solving the equation and inserting the displacement x into the equation for D gives the expression for the bulk plasma frequency (ω),

$$D= {\varepsilon }_{0}(1-\frac{{\omega }_{p}^{2}}{{\omega }^{2}+i\upgamma \omega } )E$$

(7)

where ${\omega }_{p}=\sqrt{(\frac{{e}^{2}n}{{\epsilon }_{0}m})}$ is the natural frequency of oscillation of the electron cloud (the plasma frequency), γ the damping parameter, and ε₀ the dielectric permittivity of free space.

One of the constitutive relations of linear isotropic materials with a permittivity ε is that [7]

$$D= {\varepsilon }_{0}\varepsilon E$$

(8)

On comparing the relations for D, we get

$$\varepsilon (\omega ) = (1- \frac{{\omega }_{p}^{2}}{{\omega }^{2}+i\upgamma \omega })$$

(9)

For frequencies close to ω_p, the temporal duration of damping events is quite low (approaching femto seconds and thus, close to 10¹⁵ Hz), leading to an often-made approximation that damping can be ignored, leading to [8]

$$\varepsilon (\omega ) = (1-\frac{{\omega }_{p}^{2}}{{\omega }^{2}})$$

(10)

This relation can be modified to incorporate the dielectric permittivity of the material around the plasmonic entity by establishing the relation between the dielectric constants of the two. For this purpose, the extinction of light by a nanoparticle at the resonance condition needs to be examined. The quantification of this extinction is through the calculation of the extinction cross-section (a combination of the scattering and absorption cross-sections). For a spherical nanoparticle and under the quasistatic approximation (of a homogeneous polarization or a dipolar resonance), the absorption cross-section is given by [9]

$${c}_{a}=\frac{4{\pi }^{2}{R}^{3}{N\varepsilon }_{s}^{3/2}}{\uplambda }\frac{{\varepsilon }_{i}}{{{(}{{\varepsilon }_{r}+2{\varepsilon }_{s}})}^{2}+{\varepsilon }_{2}^{2}}$$

(11)

where ${\varepsilon }_{s}$ is the dielectric function of the surrounding medium, ${\varepsilon }_{r}, {\varepsilon }_{i}$ are the real and imaginary parts of dielectric constants of the material (nanoparticle) exhibiting SPR, N is the conduction electron density of the nanoparticle, and R is the radius of the nanoparticle.

For resonance to occur and hence the maximization of absorption [10], the condition

$${\varepsilon }_{1} = -2{\varepsilon }_{s}$$

(12)

can hence be arrived at for spherical nanoparticles. Equation (10) then becomes

$$\omega =\frac{{\upomega }_{\mathrm{p}}}{\sqrt{2{\upvarepsilon }_{\mathrm{s}}+1}}$$

(13)

which intimately ties the resonance frequency with the free electron density and the dielectric constant of the surrounding material. It is to be noted here that the value of 2 is only valid for spherical nanoparticles, and for other morphologies, it needs to be modified accordingly. For example, the case of nanorods changes the value to ((1-p_f)/p_f), where p_f is the depolarization factor that is determined from the nanorod aspect ratio. In actual materials, there is often an overlap between the energies at which inter-band transitions occur and the SPR. To account for this and the fact that the damping of the resonance is not negligible and is contributed to by scattering from lattice entities (phonons, defects, impurities, other fermions (${\Gamma }_{b})$), the surface (for particle sizes approaching the mean free path of the material (${(\Gamma }_{s}))$), and radiative damping (${\Gamma }_{r})$, the relation (10) is written as [11]

$${\epsilon }_{r}={\epsilon }_{\mathrm{inter}}+1-\frac{{\omega }^{2}}{{\omega }_{p}^{2}+{({\Gamma }_{\mathrm{b}}+{\Gamma }_{\mathrm{s}}+{\Gamma }_{\mathrm{r}})}^{2}}$$

(14)

where ε_inter is the interband transition, and damping constants are given by

$${\Gamma }_{b}=^{\upnu _{F}}/_{l_{\infty}}$$

(15)

$${\Gamma }_{s}={\upgamma }_{0}+\frac{{A\nu }_{F}}{R}$$

(16)

$${\Gamma }_{r}=2hkV$$

(17)

where ${\nu }_{F}$ is the fermi velocity of electrons, ${l}_{\infty }$ is the mean free path of the conduction electrons in bulk, ${\gamma }_{0}$ is the bulk damping constant, A is a theory-dependent parameter which is dependent on the type of the scattering process, R is the radius of the metal nanosphere, h is Planck’s constant, V is the volume of nanoparticle, and k is a constant that can be arrived at from the graph relating the resonance energy of the plasmon with the measured experimental linewidth. The plasmon frequency hence changes to [12]

$$\omega =\sqrt{\frac{{\omega }_{p}^{2}}{1+{\varepsilon }_{ib1}+2{\varepsilon }_{s}}-{\gamma }^{2}}$$

(18)

where ${\varepsilon }_{ib1}$ denotes the contribution of interband transitions to the real part of the nanoparticle permittivity.

It is to be kept in mind that for obtaining the cross-sections for non-spherical particles (such as nanostars and nano dendrites), computational methods tuned to sufficiently simulate complex shapes are often necessary. Methods based on the discrete dipole approximation (DDA) such as the finite-difference time-domain (FDTD) analysis have been typically used to simulate the optical properties [13] that arise in such non-spherical plasmonic structures. FDTD algorithms solve the differential form of Maxwell’s equations by assuming appropriate boundary conditions and discretizing the space and time domain to find the electric and magnetic fields at different positions and time steps of the structure under examination.

The brief treatise in this section on the characteristics of surface plasmons enables multiple insights to be arrived at. The prediction of the plasmonic property has reached a level of maturity that lattice-level manipulations of materials and the resulting changes/induction of plasmonic virtues can be simulated knowing the equations above and incorporating appropriate boundary conditions, as will be discussed in subsequent sections. An intuitive conclusion from the comparison of SPPs and LSPs is that the latter can be much more amenable in terms of tuning compared to the former, considering the ease of excitation (due to no requirement of frequency as well as wave vector matching), the number of variables that exist for the tuning investigations, the inherent stability of discrete nanostructures than continuous films (one reason for which is the requirement of lattice matching for plasmonic films in order to have a reasonable durability when subjected to practical conditions), and the ease of synthesis (with films requiring in many cases cost prohibitive equipment). However, SPPs are worthy on their own, especially in terms of achieving resonances that are hybrids with other similar modes possible on manipulating the configurations of the films, as will be shown in later sections. It stands that the application dictates in almost every instance the choice between a film vs nanoparticle configuration, as other dictating entities such as difficulties in fabrication and compositions usually do not exist due to the rapid evolution of technology.

Plasmon-Enhanced Properties due to Nanostructured Copper

This section has been elaborated with the contextual flow of an introductory section on synthesis of CuNPs followed by the different investigations that have reported pristine as well as modified plasmonic Cu nanostructures.

Synthesis Techniques of Plasmonic Copper

A brief description of the methods and their evolutions for the synthesis of structured/morphologically controlled copper nano/microparticles and the associated changes to the LSPR properties is relevant to understanding the achievement of tuned light absorption. Simple solution-based methods such as chemical reduction [14, 15] often lead to the problems of agglomeration and the coupled loss/deterioration of the absorption properties. Oxidation in particular is a significant problem for CuNPs, with Cu₂O or CuO forming as a result depending on whether the medium is alkaline or acidic/neutral, the processing conditions, and the use of capping agents. Plasmonic properties are lost as a result due to the transition from metallic Cu to a semiconducting oxide [16]. Capping agents such as polyvinylpyrrolidone (PVP) and polyethylene glycol (PEG) can help in this regard, although differences in dispersion/agglomeration based on the choice of such agents must be kept in mind. For example, PVP has been reported to stabilize CuNPs against oxidation in a considerably better manner than PEG [17]. Techniques such as anodic stripping voltammetry can reveal the ease with which capped nanoparticles are oxidized, and analogous studies on Ag NPs which also oxidize readily report that PVP is better at preventing oxidation than PEG [18]. We have included an explicit section of how oxidation in Cu can be detrimental/prevented/exploited for and in various applications. Distinct morphologies such as nanorods and spheres, with differing degrees of sizes, oxidation, and agglomeration are possible by modifying the Cu precursors (such as Cu(DS)₂ instead of CuSO₄) and surfactant concentrations [19]. It is also possible to control agglomeration and oxidation through the use of halides such as iodine and bromine, with CuI demonstrating the best colloidal stability due to stronger attraction between Cu and I atoms and hence a more dispersed colloid (the product solubility constant quantifies this comparison between halides such as Cl, Br, and I as stabilizers) [20].

With the purpose of this review being on the different plasmonic investigations on Cu, the discussion on the synthesis modalities has been given a very cursory treatment. However, an exhaustive but incomplete summary of the shapes and sizes of Cu nano/microparticles and the methods of synthesis is presented in Table 1, and review articles on the synthesis methods for Cu [21] are also referred to the reader for a more comprehensive/specific list.

Table 1 Summary of a few morphological investigations on Cu nano/microparticles with the SPR peak locations and the synthesis method

Plasmon-Enhanced Properties from Discrete and Pristine Copper Nanoparticles

Cu Plasmons for Photoemission

Sub-wavelength features are frontrunners of the need for tuned LSPR and SPP. They have been reported to enhance multiple properties such as Raman scattering, fluorescence, second harmonic generation, catalysis etc. due in major part to the coupling between the enhanced electromagnetic field resulting from the plasmons with the phenomena responsible for the individual properties [71,72,73,74]. As an example of the probing into such features, a 100-fold increase in the charge density extracted due to photoemission of electrons from a nanostructured Cu surface has been reported. The surface consisted of patterned (through ion beam milling) nanoholes on a single crystalline Cu surface. Tuning of the absorption wavelength was pronounced while changing the spacing between the holes and much lesser while changing their dimensions and/or shapes, allowing close as well as a wide control of the absorption wavelength. Since photoemission can occur through multiple processes such as the three-step process and field-assisted emission, the Keldysh parameter was used to justify that photoemission was due to enhancements to the three-step process of electron emission from a surface rather than due to field emission. Limitations of a reduced damage threshold necessitated the reduction of the input laser fluence by five times compared to a flat surface [75]. The reason for the enhanced photoemission was postulated to be due to the efficient absorption of the incident photon energy and the transfer of this energy through plasmon decay to the ejected electrons. Such manipulations of metallic surfaces promise their usability as sources of electron beams despite their intrinsically low quantum efficiencies.

Cu Plasmonics in Energy

The concept of solar energy harvesting using nanomaterials has become increasingly relevant with time, with applications that utilize this energy including current generation, anti-reflective coatings, and heating [76]. The process of solar energy utilization through increasing carrier generation in conventionally employed materials (such as through plasmonic absorption as well as energy transfer) is a highly researched one. Briefly, the absorbed solar energy by the plasmonic oscillation is transferred to hot electron–hole pairs (in the regime of femtoseconds) from the plasmon decay [77]. Cu and Au have advantages as attractive materials in this regard as both intraband and interband transitions can be excited by visible light (the latter being 2.14 eV for Cu and 2.4 eV for Au), a major part of the solar spectrum [78]. The lower interband transitional energy of Cu generates hot electrons (inner shell electrons in the vicinity of the plasmons) of higher energy due to the close overlap. Additionally, comparing these two, Au has a higher work function (5.1 eV) than Cu (4.65 eV) [79], and Cu, therefore, can form lower energy barriers (for e⁻ transfer) with electron-accepting materials such as TiO₂. Indeed, Cu plasmonic systems in both film and NP configurations have been extensively investigated for photocatalysis when composited with TiO₂ for example [80]. Additionally, theoretical calculations of the hot carrier energies that can result from plasmon decay through direct vs phonon-assisted processes predict that the hot carriers can have a higher energy than the decaying plasmon energy due to the higher density of states (DOS) in the d bands of Cu [81]. All of these reasons make Cu a theoretically promising material for solar photovoltaics and for energy harvesting and storage applications. A detailed analysis of this comparison between Au and Cu has confirmed that the predictions are true in terms of the photocurrent generated as a result of hot carrier injection from Au vs from Cu. Cu exhibited photocurrent increases due to contributions to hot carrier injection from both the inter (550 nm) and intraband (675 nm) transition regimes, whereas Au exhibited this increase for only the plasmonic excitation due to the higher barrier Schottky barrier that the interband transition generated hot electrons could not overcome. Care must be taken however in designing the interface between plasmonic Cu and the matrix, as the distance between the plasmon-generated hot carriers (confined to near surface regions) and the electron acceptor must be at the maximum the mean free path of the hot electrons (~ 30 nm) to prevent their thermalization within Cu itself [82]. With regards to PC for energy applications, the intricacies in the processes of thermalization and scattering have been an important focal point, as summarized in Fig. 4.

Fig. 4

Summary on the investigations on PC for energy applications [83,84,85,86,87,88,89,90,91,92]

Notably, the higher imaginary component of Cu in comparison to Au and Ag at wavelengths corresponding to the dipolar resonance means that the plasmon peak is significantly damped at these wavelengths, compared in Fig. 5, adapted from the excellent work of Barbara et al. [88].

Fig. 5

a Near-field scattering efficiency of Cu-water interface for different radii of spherical particles. Extinction, far-field scattering, and absorption efficiency of Cu spheres with different radii; b 100 nm and c 22 nm in water (reproduced with permission from [88])

A tuning in LSPR location and intensity has been reported by a simple switching of the anions used for controlling solution pH between Cl⁻ and SO₄²⁻, wherein the former was able to selectively adsorb on the growing CuNPs and cause variations in the final morphologies and crystallinities. These variations led to an optimal temperature increase of 40 °C when used for LSPR-based PT heating, with a resulting promise for use as an antibacterial agent [93]. It is important to note here that the depression of the melting point of nanoparticles compared to their bulk counterparts (depending on parameters like strain, morphology, defect chemistry, crystallographic orientation of the surface, size, and shape) must be taken care of while designing them for PPT applications. Multiple studies have investigated the same, reporting significant differences from the bulk melting point of approx. 1358 K. Spheres of diameters 30 nm have their melting points as low as 450 K [94], and rods of dimensions undergo melting transitions at 673 K [95]. The PPT property can intuitively be also employed for therapy, wherein highly localized increases of temperature can kill cancerous/malignant cells [96]. The tunability of the plasmonic property is almost mandatory for such applications, as the ability to synthesize nanomaterials that absorb in the biological transparency window (650–1350 nm) is vital for in vivo therapy. CuNPs, either present alone in pertinent morphologies such as nanowires (NWs) [97] or along with a biocompatible support that helps in dispersion, have been explored with this motive. Where a temperature difference of a few degrees can be the difference between live tissue and one that can be neutralized, plasmonic heating with CuNPs can easily achieve temperature increases of 20–35 degrees [98]. A comprehensive summary of the different morphologies of Cu and Cu systems investigated exclusively for PPT applications can be found in [99].

Cu Plasmonics for Catalysis

The concept of PT heating is also of use in catalysis, termed photocatalysis, or PT catalysis [100, 101]. The obvious increase in reaction rates with temperature makes plasmonic materials that convert light into heat suitable candidates here. An important advantage of plasmonic nanomaterials in catalysis is the ability to achieve selectivity, as the energy of the hot carriers that are generated from plasmon decay depends on the plasmon characteristics [102]. All the three mechanisms of plasmon damping, viz., radiative damping into photons, Landau’s damping, and non-radiative hot electron–hole pair generation can lead to enhancements in catalytic rates. Indeed, these mechanisms have been shown in founding research wherein the photo-induced oxidative switching between Cu₂O/Cu resulted in the increase in epoxidation of propylene, an important reaction in simplifying plastic manufacturing. The energy deposition by the plasmons into the Cu₂O shell was shown to convert it into pristine Cu and restore the catalytic property. The only requirement seemed to be that of the irradiation power, where only intensities above 550 mW/cm² could cause this change in oxidation state [103]. Cross-coupling (C-C) reactions are a class of catalytic process in which complex products are formed from simpler entities. Core–shell nanocubes of Cu@Ni were able to catalyze and improve the yield by 62% of the C-C reaction of boric acid, due to the enhancement in the reaction rate provided by the absorption of light by plasmonic Cu and its conversion into heat. Compared to conventional methods of providing heat such as through a thermal transfer-based process, plasmonic heating was found to be more efficient [104]. A summary of the investigations on mechanistic aspects of PC catalysis has been compiled and presented in Fig. 6. Ag and Cu nanoparticles were compared for their catalytic enhancement rates in the oxidation of ferrocyanide, with the observation that CuNPs had an order of magnitude higher catalytic activity than Ag, yet with an accompanying limitation of a much higher oxidation-induced rate loss and leaching from the support material. Deconvolutions of the photocatalytic vs PPT mechanisms of enhancements were not performed (although possible with specialized techniques such as scanning electrochemical microscopy [105]), which would have allowed discerning the results in terms of the PT efficiency of Ag and Cu [106].

Fig. 6

An overview of different reports on the employment of PC for catalysis [107,108,109,110,111,112,113,114]

CuNPs for Enhanced/Directed Scattering of Photons: Fluorescence, Enhanced Raman Studies, and Luminescence

Since LSPR and SPP predominantly talk about absorption and scattering of light, one can intuit that the property of luminescence by materials distal to plasmonic materials can be significantly enhanced and/or altered. Properties such as fluorescence [115, 116], phosphorescence [117], and chemiluminescence [118] have been reported to be enhanced due to coupling of the fluorophores’ scattering transitions with surface plasmons and their oscillating electromagnetic field. The enhancements to fluorescence are conventionally examined through two material systems. One is where Cu alone is studied for changes in scattering as a function of particle size/environment/film thickness, roughness, and material composition. The second approach is to study the enhancements to the emission of fluorophores that are used for a myriad of applications, as a function of extremely varied parameters such as the size/shape/dielectric/composition of the Cu or the composition/distance and conjugation intensity of the fluorophore. Cu alone when present at sizes down to 12 nm exhibited an enhancement in fluorescence yield by 4 orders of magnitude (5 for 30 nm particles) compared to bulk Cu in one study [87], due in part to the enhanced local electromagnetic fields that increased the quantum yield by coupling with the incoming and outgoing photons and in part to the second harmonic generation (SHG) of LSPR [119]. Another study revealed that roughened films can enhance the emission only when the excitation energy is close to the LSPR that occurs at the local roughness features, and that the enhancement is absent for the higher energy emissions (such as the emission from the second d band of Cu centered at 3.50 eV [120]. An important observation from this study was that this absence in enhancement for the rough surface was due to the quenching of the emitted photons by the roughness features themselves, although it is probable that the absence of coupling between the LSPR (occurring at approx. 2.15 eV) and the 3.50 eV emission could be the primary cause. Enhancements to the luminescence property in particular of different PC conjugates have been compiled in Fig. 7.

Fig. 7

Investigations revealing enhancements to photon scattering using PC [121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137]

A simple but excellent illustration clarifies the reasons to the observed enhancements, as shown in Fig. 8 [134]. Multiple studies that have explored PC exclusively for enhancements to scattered Raman profiles have been illustrated in Fig. 9 along with a few unusual explorations of PC for display applications.

Fig. 8

Enhanced fluorophore interacting with a metal surface via the near-field or the far-field for the cases of region 1 (~ 50 nm) or distant (≥ 500 nm)

Fig. 9

PC studied exclusively for enhanced Raman scattering applications as well as those rare studies on PC for displays [138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157]

As shown in Fig. 10, drastic changes could be achieved in the absorbance wavelengths when PC was explored as a candidate material for displays [149].

Fig. 10

Spectral response of the nanostructured plasmonic copper film with thickness, length, and width of the slits is 30 nm 140 nm, and 60 nm respectively, and periodicity in x- and y-direction 300 nm during a reduction and b oxidation (reproduced with the permission from [148])

Plasmonic CuNPs for Non-Linear Optics

One of the manifestations of the nonlinearity in optical properties (NLO) is second harmonic generation (SHG), which is the frequency doubling of photons when passed through a non-linear material and has huge potential, for example, in realizing high-energy lasers, specifically those with tailored wavelengths (instead of only the typically available wavelengths that are limited in terms of their length by the material’s electronic structure). Local field enhancements are a crucial reason for the enhancements to both SHG and SERS, and measurements of the same can be helpful in comparing and choosing the appropriate plasmonic material of choice. In this regard, the drastic drop in field enhancements for excitation wavelengths below approx. 650 nm (due to an increased damping of the resonance from inter-band excitons) necessitates this as a limit for plasmon-enhanced SHG from Cu [158]. Of the very few studies that have explored the NLO potential of Cu, that of PVP-protected CuNPs revealed third-order nonlinear susceptibilities equivalent to that of AuNPs, leading to their prospective potential in four-wave mixing. The plasmonic component being the reason for this observation was not discussed; however, it is probable that the LSPR excitation played a part here as the excitation wavelength used close to that of the LSPR of CuNPs [159]. From the myriad of investigations that report the possibilities of non-linear SHG on multiple plasmonic materials [160,161,162,163,164,165], it is proposed that the potential of Cu for the same is at a very nascent stage of study.

Cu Plasmonics for Sensing

Sensing with CuNPs alone without a protective layer presents problems in terms of oxidation, as will be discussed in subsequent sections. However, different architectures incorporating Cu in the nanometer sizes have been explored regardless. An example is the use of thickness-controlled deposited Cu shells on a SiO₂ island, with the prediction of a higher refractive index (compared to air) and bio-compatible SiO₂ being more suitable for a red-shifted and stable plasmon. The influence of thickness was detrimental post 30 nm, for which the highest intensities of absorption were observed. The use of these capped Cu shells for DNA sensing was investigated, and sensitivities down to 10 fM were reported [166]. A general comment here is that for sensing with CuNPs formed from film deposition and annealing, the relation between the thickness of films and the observed plasmonic attributes such as intensities, positions, peak widths, and stability is not always universal, as other factors such as packing density (which controls light penetration) and surface faceting of the islands of Cu that are finally achieved also play a role. CuNPs have been investigated for colorimetric sensing as well, with the ability to detect sulfide ions at concentrations between 12.5 and 50 µM. Interestingly, the oxidation of the synthesized nanoparticles did not happen for a few days, attributed to being stored in airproof vials and the protective nature of the CTAB ligands used during synthesis [167]. These are among the few studies that have explored the potential of discrete nanoparticles of Cu for sensing, more investigations being limited mainly by the tendency for oxidation and probably difficulties in achieving unique morphologies that can be better in terms of the plasmonic properties compared to conventional spherical ones.

We have so far presented an overview of the different purposes for which Cu has been explored in the nanoparticle form for plasmonic applications and explorations. The other construction of plasmonic Cu is the Cu film, wherein interesting plasmonic phenomena can be invoked apart from the traditional SPP wave. The different applications, limitations, solutions, and outlook on the use of Cu films exploiting in particular the SPP characteristics are discussed in the subsequent sections.

Copper Plasmonics in Films

Plasmonic Cu Films for Sensing

The field of sensing is a prime area of application for plasmonic films, due to the inherent interface sensitivity of the plasmon to charge interactions. Films of Cu have also been explored in this regard, although with the consideration that a protective dielectric be deposited on the film to avoid oxidation. With this ideology of fabrication, films of Cu protected by SiO₂ or Al₂O₃ with a graphene oxide linker have been reported for studying neutravidin-biotin interactions. The refractive index sensitivities for the Cu films with these layers were found to be 55% and 75% higher than for even Au films with no dielectric on top, attributed to the significantly lower optical losses in the energies investigated for Cu compared to Au [168]. The Langmuir–Blodgett films of copper phthalocyanine have been shown to be a potential sensor material for NO₂ [169]. An issue with multi-elemental protective films is their deterioration over time and use due to the high probability of the presence of defects including pores at the time of deposition, which can serve as diffusion paths for oxygen to access the underlying Cu, oxidizing it over time and eliminating the plasmonic effect. Close-packed and conformal films can hence be inherently better at protection, and graphene has been attempted as a protective layer with this reasoning on Cu films. The atomic thickness with which graphene can be worked with aids in reducing the extent of shielding of the plasmonic field compared to the thicker layers of the multi-elemental materials as well, aiding in enhanced sensitivities than other thicker films. The defect density in graphene is again vital here, and graphene grown elsewhere and transferred onto Cu was better at protection than in situ grown graphene due to a higher defect presence in the latter films, arising from surface effects from lattice mismatch during film growth. The resulting protected Cu films were promising for the sensing of graphene hydrogenation, as well as waveguides for light transmission [170]. With this investigation as proof, a subsequent study has revealed the prospect of this Cu/graphene layer composition deposited on a photonic fiber as an effective sensing configuration with a sensitivity of 2000 nm/RIU. An increase in sensitivity was achieved through intentionally incorporating appropriately sized holes (300 nm diameter) on the fiber surface prior to Cu/graphene deposition, leading to the coupling between the SPP mode and the LSPR of the holes. The number of layers of graphene had a profound impact on the sensitivity, as more than two layers led to a significant damping of the SPP. Due to the attenuation of the SPP into the metal being much stronger than the dielectric, a higher refractive index dielectric as an intermediate layer can be intuitively understood to be better for sensing due to the resulting improved accessibility of the plasmon field to interactions happening on the surface. With this understanding and with regard to improving the sensing performance of Cu films, materials amenable to deposition as a continuous film with a high refractive index and a low thermo-optic coefficient along with minimal lattice mismatch can be ideal. An investigation of the use of III-V nitrides which have a low thermo-optic coefficient, similar lattice configuration to graphene (hexagonal) high refractive index, and a near-zero extinction co-efficient (and hence unwanted incident light attenuation) at the Cu SPP wavelength of 633 nm has proven to enhance the sensitivity of simpler Cu/graphene films with this reasoning [171]. The theoretical study revealed that the higher optical quality (judged from obtaining the best evanescent field enhancement and sensitivity) of InN among the other nitrides resulted in the best performing composite film.

Native oxides of Cu have also been investigated as the easiest to form protective layer for Cu in sensing applications. A study on the best thickness of Cu and an intentionally created oxide layer (through annealing) revealed that for the maximum sensitivity of the film when the refractive index was changed between that of air and water, the optimal thickness of Cu was 44 nm and that of the oxide was 1 nm for a film deposited on a glass prism [172]. A similar investigation in terms of the sensitivities between air and aqueous phases but with the objective of comparing the sensing configuration has been performed, in which the configurations of Cu films deposited on prisms, those with periodic corrugations (the grating configuration) and those deposited directly on optical fibers were fabricated. Results revealed that the prism configuration exhibited the best sensitivities, but that this was primarily due to difficulties associated with the other two configurations, viz., the inhomogeneity of films for the fiber configuration and the loss of a significant portion of the incident light due to diffraction in the grating configuration. A thin layer of Au was deposited to prevent oxidation of the films [173]. A similar finding of a lower refractive index sensitivity, when films in the Kretschmann configuration and those with a grating configuration were studied, has been reported, wherein Cu was found to be more sensitive than an Au film. The reason for the grating configuration being an order of magnitude less sensitive could be due to inefficient evanescent wave coupling to the plasmon, as the evanescent wave is not entirely restricted to the metal/dielectric interface in this configuration [174]. Refractive index sensitivity studies with water and glycerol with Cu films prepared through chemical deposition and protected by a layer of benzotriazole molecules led to values of up to 131 nm/RIU depending on the film deposition parameters [175]. The choice of materials that can improve the sensitivity is more critical with regards to sensing with films rather than the film parameters themselves, such as protecting/permittivity enhancing over layers and the use of adsorption enhancing molecules, as revealed by a study employing Cu films with Fe₂O₃ and antimonene over layers on Cu configured in the Kretschmann geometry. Fe₂O₃ improved the permittivity, and antimonene was used for a more efficient binding molecule due to its highly customizable functional group chemistry. Refractive index sensitivities of up to 398 nm/RIU were reported from the simulations [176].

Cu Films as Plasmonic Waveguides

It is of note here that the loss of the Cu plasmon wave is significantly dependent on the surface quality (against impurities) and surface crystallinity (against a high concentration of grain boundaries) as well, significantly impacting the propagation length which can reach values of up to 65 µm at 750 nm for oxide-free but polycrystalline films [177]. These values are of significant importance in the field of subwavelength optical guides, of which plasmonic materials are deemed as frontrunners. Cu is especially relevant for wave-guiding applications due to its compatibility with existing CMOS processes. But Cu by itself is not preferred for wave-guiding among the coinage metals due to lower propagation lengths stemming from high ohmic losses, albeit having a smaller mode confinement volume [178]. However, lengths > 40 µm at the telecommunication wavelength of 1550 nm for 170 nm thick films protected by SiO₂ and SiN layers have been observed [179]. Of importance here is that the propagation length is not only governed by the material but also by the construction of the waveguide.

Waveguides can be made significantly more efficient in terms of propagation lengths by sandwiching a low-permittivity material between the metal and the conventional dielectric as the propagating mode is pushed out of the metal into the sandwiching dielectric [180]. Interesting to note here that the selection of the intermediary dielectric is opposite here (with regards to permittivity) than when one is selected for sensing. The resultant coupling between the SPP mode in the metal-low permittivity dielectric and the fundamental mode in the higher permittivity dielectric leads to a strong confinement (within areas down to 1/100^th of the diffraction-limited area for the wavelength) of the propagating hybrid mode, significantly enhancing the propagation distances [181]. A common configuration of this hybrid plasmonic waveguide is Cu-SiO₂-Si [182,183,184], which has also been reported as the first experimental configuration to demonstrate polarization splitting in the hybrid waveguide configuration. This splitting results in the TM excitation being able to propagate along the device, whereas the TE is not [185]. The same configuration has also been demonstrated as a compact electro-optic modulator with very low waveguide insertion loss [186]. Cu-TiO₂-Si was a configuration investigated as a waveguide-based resonator due to the negative thermo-optic coefficient of TiO₂ [187]. Spoof surface plasmons, waves propagating in the 1 × 10⁹–11 × 10⁹ Hz regime, have been realized with staggered waveguides made of Cu films. Propagation distances over 150 times the wavelength have been reported with features that were in the form of squared bars, and the distances depend on the staggered separations between the bars [186].

Cu Films in Non-Linear Optics

Films designed for non-linear optics are of benefit for example in making neutral density filters, as lasing materials, as emission sources of up converted photons, and as other light manipulators. Cu films exhibiting one such NLO characteristic of reverse saturable absorption (RSA) have been fabricated through pulsed laser deposition. RSA is the increase in absorption of the excited state electron population in comparison to the ground state electron population, which is notably contributed to not only by plasmon absorption but also by inter-band transitions. Manifestation of this behavior was found to be closely associated with the particle size in one study, which for sufficiently small particles related directly with the hot carrier generation efficiency from plasmonic decay. The line between calling the deposited structures films vs particles was very thin in this study, as although the deposition was described as a film, the plasmonic behavior (observation of a broad absorbance) was attributed to discrete and polydisperse nanoparticles [188]. The standing of a particular material composition/configuration for non-linear RSA applications is quantified by the non-linear absorption coefficient, which was on the order of 10⁻⁴ cm/W for these laser-prepared CuNPs. Other studies have examined different compositions and configurations, and the analogous coefficients were 2–6 × 10⁻⁵ cm/W for Cu-SiO₂ films [189], 0.01–1 × 10⁻⁸ cm/W for ion-exchanged Cu nanocomposite glass [190], and 10⁻⁶–10⁻⁷ cm/W for Cu implanted silica glass nanocomposites [191], demonstrating the possibilities for such materials for optical attenuation in high power lasing applications for example. Towards another investigation of NLO from Cu, an unconventional and novel study on the excitation of tip-induced plasmons (TIP) due to energy transfer between electrons tunneling between the tip of a scanning tunneling microscope (STM) and a Cu phthalocyanine (CuPh)/Au film has reported an enhancement in the emission of up-converted photons from this configuration. An enhancement in the inter-band electronic transition-induced photon emission at a wavelength of 720 nm from the CuPh was observed. The up-conversion was confirmed from the fact that the bias voltage of the tunneling electrons (< 1.7 V ≈ 720 nm) was lower than the energy of photons emitted (720 nm) [192]. The use of nanopatterned Cu films as cathode sources for electron emission has been reported. The enhancements of such a surface composed of spherical holes (and localized resonances hybridizing with the SPP mode) to the electron yield was 1200 times that of a flat surface which does not exhibit localized resonances. The emitted electrons had a narrow energy distribution (judged from the value of the Keldysh parameter), and the enhancements originated from SP assistance, confirmed through simulations and calculations of expected enhancements assuming the changes occurred only due to a lower reflectivity of the patterned films (the Fowler-Dubridge theory was used for quantifying the enhancements due to lower reflectivity). The angular spread, emission time scales, and the damage threshold of laser fluence of the nanohole array cathode were comparable to the most efficient metallic cathodes [193].

Cu Films for Enhanced Scattering

Scattering of light by plasmons being a dominant field of study and Raman scattering being a major benefactor from plasmonic enhancements, Cu films have also been explored along these lines, albeit to a much lesser extent than Au and Ag due to the high oxidation instability and often comparable or inferior enhancements to the latter two. SERS of Cu phthalocyanine molecules deposited on Cu resulted in a 23-fold enhancement when excited at the SPP wavelength of 632 nm of Cu, compared to the 50-fold enhancement observed for Au. The collection efficiency of the scattered light from the SPP was improved using a Weierstrauss prism and in the Otto configuration (the latter for control over the angle of reflectance) [194]. The coupling of propagating SPPs and LSPRs can be very interesting due to the intense electromagnetic fields concentrated at the interfaces between the NPs and the films. In this regard, islands of Ag with their LSPR have been investigated as a coupled system with the propagating SPPs of Cu films for enhancements to the fluorescence of dye molecules. Both emissions of photons into free space by the proximal dye molecules of rhodamine, as well as plasmon-coupled fluorescence (PCF-detected at the back of the Cu films), increased with the presence and increasing coverage of the Ag islands, and at coverages approximating Ag films instead of NPs, the enhancements diminished considerably indicating the beginning of the loss of the metal-enhanced fluorescence property. A narrowing of the emission angle of the PCF was also a consequence of the presence of the Ag islands [195]. Similar observations were made in one of the other rare investigations of MEF on Cu films, with 40 nm thick films exhibiting the highest increase in PCF by 1.39 times of sulforhodamine fluorophores [196]. While it is intuitable that such films may not stand the test of time due to their oxidation-prone nature, it is probable that thin protective layers such as graphene as discussed earlier in this article can be a viable approach for realizing durable and plasmonic fluorescence enhancing Cu films.

SERS being dependent on the hotspots created and the extreme field localizations at these locations, films of plasmonic materials empowered with surface roughness and/or roughness features such as other nanoparticles can be interesting. The role of the hotspot in SERS studies (which can also be used to monitor scattering from molecules created at the hotspot assisted by plasmons) can be confirmed for example from two separate studies. The first used a film of Cu on which Ag nanoparticles were and were not present on the surface, for which the latter morphology catalyzed the reaction (the dimerization of 4-nitrobenzenethiol) and the former did not on the excitation of SPP/LSPR, respectively [143]. The second study was also on photocatalysis (for the conversion of 4-aminothiophenol to p,p′-dimercaptoazobenzene) but employed CuNPs which could readily cause the reaction [197].

Optical Nanoparticle Trapping by Cu Plasmons

Optical trapping is the phenomenon of exertion of forces on an agent (example: an NP) for its manipulation by virtue of its interaction with light. A coherent/concentrated EM field is required for this to be possible, such as from a laser. Plasmons intrinsically can be of applicability in this field and have been explored for this property. The tuning of the separation distance and geometry of one such Cu plasmonic nanoantenna for the manipulation of polystyrene nanoparticles resulted in the intense EM fields exerting forces up to a few femtonewtons on them. It is important to note here that the enhancements to the EM fields that resulted in the trapping were due to excitation at 1064 nm and not the plasmonic wavelength of Cu, which should be equal to if not better for optical trapping if investigated due to the inherently magnified and localized fields during resonance. The potential use of polystyrene for biological applications led to the choice of this wavelength for the study, wherein in vitro applications can benefit from this manipulation [198].

On the Oxidation of and Changes Thereby to the Plasmonics of Copper

The Nature of Oxidation of Cu

Many theories and supporting evidence exist for the highly detrimental (from a plasmonic view) phenomenon of metal oxidation [199, 200]. However, we restrict the discussion on this topic to reports specific to the observation and investigation of oxidation to plasmonic Cu films and nanoparticle structures. Copper has an extremely high tendency for oxidation, with the dominant oxidation product being Cu₂O [201] and also CuO (with intermediate phases such as CuO_x where x = 0.67 [202] present). It is to be noted however that a range of stoichiometries of CuO_x is possible depending on temperature, film porosity, thickness, and particle diameter with X values from 0.67 to 1 [203, 204]. The growth of the oxidized layer is linear depending on the ambient and the nature of the film/NPs, at an oxide film formation rate of 0.004 nm/day when stored at ambient temperatures (23 °C) and atmospheric pressure for example [205] (contradicting numbers exist here for the oxidation rates depending on particle sizes and ambient conditions such as the rate of 0.034 nm/day for 100–140 nm spherical CuNPs [206]). These numbers predict a continuous growth of an oxide film, whereas other reports observe a saturation in oxide formation to the tune of 13 nm thick films after a few hours irrespective of oxygen pressure. Multiple mechanisms have been proposed as to the oxidation mechanism such as the uniform passivation of the CuNPs surface by oxygen to form the oxide [207] which was superseded later by the observation where oxide islands nucleate on the surface and coalesce to form the film [208]. Nevertheless, the quick oxidation of pristine Cu with a resulting layer not exceeding a thickness of ~ 15 nm (composed of Cu₂O) followed by the occurrence/conversion of other species such as CuO and CuO_xH₂O [209] is the inference from research on this subject, post which further diffusion of oxygen into the oxide becomes a limiting step for oxidation of the underlying pristine Cu. The self-limiting nature of this process (when exposed to air for example) has been reported to disappear even at 50 °C, post which complete oxidation of NPs occurs [210]. The resulting oxide is composed of mostly Cu₂O with CuO moieties present in proportions that depend on temperature. Although the amount of formation of Cu₂O is largely near the surface of the bare metal [211], significant differences can occur in the plasmon oscillation as this is a phenomenon confined to the surface, leading to a gradual loss of the property due to a decrease in the free electron density. This loss and changes to the plasmonic properties of Cu in general have been studied in detail by multiple researchers [212,213,214]. Conflicting reports thus exist in this continuing field of CuNP oxidation research as revealed by the discussion so far, to be added to the prediction and confirmation (through theoretical modeling of the barrier energy for continued Cu diffusion for oxidation) of a maximum self-limiting oxide shell thickness of only 0.56 nm. Although, in this letter study, theoretical calculations considered the various barriers to diffusion of Cu atoms at the core towards the surface, it was observed that for shell thicknesses up to 0.56 nm, the barrier to Cu diffusion from the bulk to the surface was low enough for oxidation to proceed [206]. It stands that the phenomenon of oxidation while considering Cu fluctuates highly in terms of the time required and the extent of self-limitation depending on factors such as purity, morphology, temperature, and surface characteristics (such as crystallinity, roughness, porosity, etc.) and the ambient environment (with its own decisive parameters such as pressure and nature of the oxidizing species) [215] but cannot be unaddressed unless stabilized or isolated against even the mildest of oxidizing environments.

Prevention/Reversal/Utilization of Cu Oxidation as Relevant to Plasmonics

Reversal

Oxidation in Cu can be prevented/reversed by a few procedures. For example, the treatment of oxidized Cu with acetic acid was found to remove the oxide layer through the formation of cupric acetate and a resulting return of the plasmonic metallic Cu [57, 216]. There was no loss of the thus exposed Cu due to etching, rendering it safe for such procedures. Adoption of appropriate post treatment procedures such as drying with inert N₂ instead of a water rinse/air drying is essential [217], to avoid the reformation of the various oxide species. Although these species reappear once exposed to ambient conditions, this method presents a solution where devices based on plasmonic Cu can be prepared where vacuum transfer between fabricating equipment can minimize this re-occurrence. If avoiding hydrated environments is not a possibility, vapor phase etchants such as 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (H + hfac) can give similar results of removing the oxides [218]. The oxidation can also be reversed for example by subjecting to a strongly reducing atmosphere such as H₂ where all the phases of oxide were observed to change into elemental form. This study has investigated only CuNPs and not films, and hence, the redox behavior of films can be expected to be different based on porosity, thickness, etc. [204]. Yet another method of reversal seems to be direct irradiation with the light of oxidized Cu. With regards to deposited Cu nanoclusters, the facet design of the clusters can be useful in prevention. Preparing clusters through gas phase aggregation such as through physical vapor deposition leads them to typically have a monocrystalline nature {(111)}, with these facets having a higher coordination compared to polycrystalline NPs. This results in fewer reaction sites for oxidation. Combined with a process such as plasma ozone treatment which resulted in the formation of a Cu (II) oxide compared to Cu (I) and where the Cu core-Cu (II) shell was found to be much more stable against further oxidation than when Cu (I) forms (typically when exposed to ambient conditions during storage), this methodology of CuNP production can be a prospective solution to the oxidation problem of plasmonic Cu, as shown in Fig. 11 [219].

Fig. 11

LSPR band for large CuNPs at different times a kept in ambient air, b kept in a nitrogen atmosphere, and c ozonated and kept in ambient air (reprinted with permission from [219])

Prevention

In terms of prevention, the obvious choice of stabilizing ligands is the most straightforward solution. Multiple ligand types and classes including thiols and acids have been tested for their effects on stabilizing CuNPs against oxidation, with thiols and oleic acid being those which lead to oxidation stability while simultaneously avoiding aggregation [220]. For thiols, longer chains resulted in more stability to oxidation but were also found to cause a rapid loss of plasmonic Cu due to thiolate formation with the help of atmospheric oxygen. This observation was not found with oleic acid, wherein the double bond in the oleyl chain was responsible for stabilizing the CuNPs [221]. Thiolate immobilization (which is usually self-assembled) can thus prevent oxidation, but probably at the cost of the plasmonic property of Cu [222]. The chain length and the composition of the thiol affect the stability provided against oxidation, with longer chains stabilizing better due to the reduced permeability of oxygen through the chains [223,224,225]. One-step synthesis techniques such as irradiation by high-energy γ rays of the matrix in which CuNP precursors are dispersed are also effective ways to realize NP synthesis (by ionized species generated by the irradiation) and to avoid oxidation simultaneously (as the dispersion matrix, PVA, is impervious to oxygen) [226]. Stabilizing ligands which contain multiple moieties that can bind to the CuNPs can be found in molecules such as polyethyleneimine, allowing enhanced stability and an alternative pathway to prevent oxidation compared to single- and double-molecule-capped CuNPs [227]. Prevention of CuNPs from oxidation to preserve SPR by various modalities has been reported in brief in Fig. 12.

Fig. 12

Research on the prevention of PC oxidation towards LSPR preservation [228,229,230,231,232,233,234,235,236]

Utilization

The oxidation of CuNPs leads to interesting effects as well due to the inherent nature of the LSPR property. For example, the commencement of the process does not immediately quench the plasmon oscillation, until the coupling of the photon to the valence electrons proximal to the surface is no longer possible due to the screening of the incident photon field from them [237]. In fact, depending on the nature of the solvents (i.e., whether they are capable of forming electron-sharing associations like pi- bonds), the plasmon intensity has been reported to increase if the solvents where the CuNPs are dispersed change from one that is not capable of pi-bonding to one that is. The eventual result is the attenuation of the plasmon peak and the appearance of the exciton peaks, depending on the rate of oxidation [238] in such solvents. Yet another prospect is the use of oxidized Cu as an electron scavenger, preventing recombination and leading to the generation of holes for catalyzing reactions. In a study with ZnO composited with Cu, the oxidized phase of Cu (CuO), which resulted very quickly on exposure to the ambient, was used as an advantage to scavenge electrons on photoexcitation of ZnO, wherein these electrons were used for reduction of the CuO, whereas the holes were used for catalyzing production of H₂O₂. For such an application, Cu is most beneficial compared to Au and Ag due to its lower half-cell potential (0.52 V) compared to Au (1.69 V) and Ag (0.80 V) [239]. A similar enhancement to the catalytic activity of Cu@CuO also points to the usability of an oxide shell, for catalysis at least [240]. Such studies reveal the prospect of achieving tailored absorption through facile manipulation of the CuNPs’ environments and can be useful while trying to achieve NPs with a better shelf life for various applications.

From the discussion thus far, the phenomenon of oxidation in Cu need not always be viewed as a detrimental process. The intricate control of the parameters such as the proportion of the CuNPs that are oxidized, the thickness of the oxide layer, and its use for applications as pointed out so far leads to the postulation that Cu plasmonics can lead to a fruitful collaboration with oxidizing agents in terms of realizing specific goals for pertinent applications.

Plasmonics in Doped/Non-Elemental/Composite Copper Nanosystems

Pristine plasmonic nanomaterials gradually lose their plasmonic property on increasing the particle size, due to increasingly stronger damping of the plasmon with size. Moreover, the plasmonic properties of pristine metals are locked-in post deciding their size, shape, and composition. However, doped nanomaterials can have significantly different LSPR manipulative abilities, due to the possibility of making changes to the stoichiometry (and by extension the dielectric constant/free electron density) by doping, for instance. Apart from doping, temperature modulation (including thermal aging), phase transitions, oxidation, and field-assisted copper migration can also tune the LSPR of Cu [241, 242]. Doping can be achieved predominantly through two modalities, viz., inclusion of a different element and by self-doping [243], wherein the parent molecule is imbibed with vacancies to change the stoichiometry. Copper is particularly amenable to creating non-stoichiometric doped compositions due to its low chemical potential which leads to its release from the solid phase relatively easily, leading to the induction of LSPR due to self-doping. The former modality for inducing LSPR through doping/alloying is the compositing of pristine plasmonic materials with other materials such as semiconductors [244] and polymers with marked improvements in the resulting properties. Herein, we attempt as much a comprehensive discussion as possible on the different classes of Cu-based plasmonic nanomaterials, with a focus on the plasmonic aspects.

Copper Chalcogenides

Chalcogenides of Cu have been the major players among the different plasmonic Cu compounds, being extensively investigated for their plasmonic properties. Hence, the section on chalcogenides alone has been divided broadly into their origins and their applications.

Plasmonic Origins and Virtues of Chalcogenides and Their Applications

Summaries treating Cu chalcogenides as sub-sections of a broader perspective, for example, on doped semiconductors and metal oxides, as well as summaries of chalcogenides being explored for different applications such as theranostics and catalysis, have been published [245,246,247,248,249,250]. Broadly, the investigations on plasmonic copper chalcogenides have been centered primarily on Cu₂S and Cu₂Se, with compounds such as Cu₂Te and CuSSe also being investigated with a decreasing order of frequency. It is of note that stoichiometric chalcogenides are not plasmonic due to the absence of sufficient free carrier density, while Cu-deficient compounds are plasmonic due to the presence of holes. Rigorous reviews on plasmonic Cu sulfides treat these compounds in terms of multiple viewpoints such as their compositional integrity in various situations, the limits of plasmonic manipulation, the applications where they can offer benefits, and the property enhancements possible by doping multi-valent ions into their lattice [251]. Their applicability to the control of solar radiation has also been widely investigated (through films deposited via chemical bath deposition on multiple substrates) as efficient reflectors of infra-red radiation while permitting visible light penetration [252,253,254]. On this perspective, only their light interaction properties of transmittance, reflectance, etc. are quantified, with the mechanisms behind the observations either attributed to their semiconducting nature or not discussed. Considering the focus of this work, only research that has discussed and discovered aspects in the plasmonic perspective and those that explicitly acknowledge/treat this perspective of Cu_xS have been included.

With regard to the plasmonic Cu_2-xS family, these nanomaterials have had an interesting history of being identified as plasmonically active nanomaterials in the infra-red from being wrongly assigned as indirect band gap materials [255,256,257]. The validation of the LSPR property through assessment of the optical absorbance dependence on free carrier density (for Cu₂S) was the point from which plasmonic chalcogenide research took off [258, 259]. Of note is the fact that the confirmation of the LSPR property in the first study derived from comparisons of the observations of the optical absorbance of stoichiometric Cu₂S which did not exhibit LSPR absorbance [260]. Since then the investigations of this family have grown in terms of manipulating LSPR through various means, although in many instances continued attributions to the d-d transitions instead of plasmonic excitations in these materials [261,262,263,264] are made. The composition of Cu_2-xS can accommodate large variations in Cu vacancies, leading to structures such as chalcocite (Cu_1.997-2S), djurleite (Cu_1.93–1.97S), digenite (Cu_1.8S), and anilite (Cu_1.75S). As cursorily mentioned earlier, LSPRs of non-stoichiometric Cu₂S are mediated by holes rather than electrons as they are the free carriers in this p-type molecule and are generated when CuS gets oxidized to the different non-stoichiometric forms. The relatively lower free electron densities that result lead to LSPRs in the NIR region, with high tunability in their position and bandwidth depending on free carrier density, shape, and the refractive index of the surroundings (such as stabilizing ligands/shells/solvents). To this extent, different shapes and sizes of Cu_2-xS nanomaterials have been developed, including disks [265], spheres with precise manipulation and reversibility of the SPR property (through the reversible processes of oxidation in air /reduction with tetrakis hexafluorophosphate) [266], and spheres with biasing based modulation of the free carrier density [267]. It is worth noting that the LSPR of these chalcogenides can be controlled more drastically by controlling the transformations between the different phases rather than by controlling the morphology, and even then, within a phase, only a limited variation in non-stoichiometry exists. It is also established that the barrier for further transformations into other phases is higher than that for achieving variations in stoichiometry within a phase. Hence, by employing oxidizing conditions with suitable potencies, fine tuning of the LSPRs of chalcogenides can be possible [268]. A very useful equation that allows prediction of the LSPR energy (${\omega }_{sp}$) is given as follows [269]:

$${\omega }_{sp}=\sqrt{\frac{{\omega }_{p}^{2}}{1+2{\varepsilon }_{m}}-{\gamma }^{2}}$$

(19)

The terms ${\omega }_{p}$, ε_m, and γ are as explained in the previous sections. It is of importance to note here that the above equation is valid when the LSPR is separated from the band gap excitonic transitions. It is important as well that the linewidth γ is influenced by variations in vacancy density as well as the particle size distribution and that estimating the value of the hole effective mass in the calculations of ω_p is not always performed for the purposes of validation of the calculations. Accurate values of hole effective mass which can be arrived at from the valence band curvature are very important to be able to predict the location and the width of the resonance for hole-based LSPR materials such as chalcogenides. Indeed, substantial variations in the effective masses exist for the CuS family, ranging from 0.55 m_e for CuS to 0.8 m_e for Cu_1.96S [270, 271]. With regard to how many free carriers are required for a plasmon oscillation, such estimations for a 6 nm QD of Cu_2-xS, for example, lead to a figure of approximately 200, ample enough to sustain LSPR which was confirmed by spectroscopic investigations. Care must be taken while dealing with interpreting the LSPR behavior of Cu_xS, as multiple studies have also revealed the possibility of Cu existing in monovalent or divalent states depending on the lattice coordination environment. It can be predicted then that LSPR tuning is possible for the latter than the former, and it is evident that the full potential of LSPR manipulation of CuS/Cu₂S is not always addressed [272, 273]. Indeed, the conversion between covellite (CuS) and chalcocite (Cu₂S) through reduction with Cu⁺ ions is one of the transitions that have been well documented, although the manifestation of this transition on plasmonic virtues such as the study on band structure and temporal stability leaves room for discovery [274].

Since Cu is inherently unstable unless protected by various means as discussed in earlier sections, investigations on protecting CuNPs and films against oxidation while still maintaining the electronic structure of the free carriers have been a continuing endeavor. A promising but less explored route of stabilizing the hole-free carriers, not by Cu vacancies, but by using anions adsorbed on the surface, has been realized in this approach. The compound of ferrocenium triflate was found to be a suitable valence hole stabilizing species by adsorbing as an anion on the surface while still preserving the Cu stoichiometry. This approach was, however, quantitatively less effective at stabilizing the holes (by approx. 300 mV) compared to the conventional approach of creating Cu vacancies to realize free holes. Interestingly, both modes of realizing free holes resulted in indistinguishable LSPR bands, indicating the suitability of this approach where Cu deficiency leading to phase transitions is not desired [275]. Plasmonic probes of reactions can be very sensitive to the refractive index (n_m) according to the following equation [276]:

$${\varepsilon }_{r}\left(\upomega \right)= -2{n}_{m}^{2}$$

(20)

They can thus be used to monitor the reaction environment to determine parameters like the pH, charged species present and in general to ascertain the redox conditions. Specifically, the activities of oxidants such as alkylamines and iodine promote the creation of Cu vacancies whereas reductants like thiols and sodium biphenyl quench them, causing a blue and a red shift of the LSPR respectively, proving that the Cu plasmon can also serve as an indicator of the species present if the reverse approach (of interpreting the shift in LSPR as an indicator) is followed. Additionally, the mobility of Cu in CuS has been proven to be high and essentially without restriction (due to multiple theories one of which is the weak-electrolyte model) [277], thus not factoring as a rate-limiting step in these reactions and enabling them to be used as prompt probes of the reaction conditions [276].

Cu Chalcogenide Sulfides

The concept of PT catalysis due to plasmon-induced heating has also been investigated with plasmonic Cu chalcogenides. Anilite (Cu₄S₇) has been used with this motive for the cyclocondensation of 1,3-cycloheanedione and 3-methyl-2-butenal. The porous nature of anilite achieved helped in maximizing the absorption in the infrared, leading to enhancements of approximately five-fold of the catalytic rates. The crevices of the pores were also predicted to contribute to this enhancement by facilitating heat transfer through their increased surface area [278]. Analogously, enhancements to the catalytic reactions of the Suzuki coupling reaction, the oxidation of benzyl alcohol, and the hydrogenation of nitrobenzene have been reported to be due to the plasmon-enhanced performance of anilite@Pd nanoparticles. The infrared absorption of anilite compared to the UV–visible for Pd caused only anilite to be the contributor to the increased catalytic rates at NIR wavelengths. Separation of the contributions to this enhancement due to thermal heating (due to plasmonic absorption) and due to the photocatalytic channel was done through a simple experiment in the presence and absence of light (while still providing external heat when illumination was not present). The considerably reduced catalytic rates for the latter condition confirmed the LSPR route for photocatalytic enhancement. Corroboration with work function calculations leads to the mechanism of hot hole injection (due to LSPR-induced e–h pair generation) into Pd from Cu as the cause for this enhancement, whose resulting p-type nature served as the initiator for all the reactions such as the dissociation of H atoms [279]. The observations here support the theoretical band structure calculations for Cu from other studies, which report generation of hot holes with a considerably higher energy than hot electrons (by ~ 2 eV) and hence an easier injection into other materials, making them promising for such applications. This asymmetry in energy distributions of hot electrons and holes makes Cu an efficient material for hole injection and an attractive prospect in plasmon-driven hot carrier research [280]. Yet another hydrogenation reaction in which furfural was catalyzed by Cu-LSPR-mediated dissociated H₂ into furfuryl alcohol reports hot carriers as the energy source for the dissociation. The encapsulation of CuNPs in a thin film of carbon enabled oxidation-stable photocatalysis during the hydrogenation process [281]. The easier migration of free holes within plasmonic Cu_xS through creating this material with the help of a Cu_xS seed vs without has been studied. The observations indicate that the seed-based approach led to a much higher concentration of Cu vacancies and, hence, a better photocurrent density than the starting material of stoichiometric CuS. The increase in free hole density by approx. 2.5 times between materials achieved through these two processes was sufficient to cause a shift by approx. 1000 nm in the LSPR wavelength [282].

Theranostics, the field combining both branches of medicine of therapy and diagnostics, has evolved in a considerable manner with the advent of nanotechnology. In this regard, PT therapy of tumors has been a highly explored one, especially with plasmonic nanoparticles due to their widely tunable morphology and stability. The PT property of plasmonic digenite was found to be effective for cell destruction due to its plasmonic absorption in the NIR region. Comparisons with AuNRs led to the conclusion that they can be superior to the latter, although comparisons with optimally designed AuNRs for PT therapy were not done [283]. Materials with more than just the PT property, such as those that can also enhance the imaging component of the therapeutic methods, can be intuited to be better than those capable of a singular function. In this regard, with particles between 10 and 50 nm being ideal to have the best retention times in vivo, magnetic Fe₃O₄@Cu_2-xS nanoparticles with synergistic advantages of imaging (magnetic and thermal) and therapy (by infrared absorption induced heating and ablation) of tumors have been reported [284]. Regardless of the controversy surrounding the nature of the excitation of Cu chalcogenides, such composites with both magnetic and optical control have immense interest stemming primarily from their broad absorption, allowing exploitation of their absorbing nature at both the biological windows (808 nm and 1064 nm). The efficacy of the lower energy window was much higher than the higher energy, and the molar extinction coefficient was the reason. Specifically, the higher extinction coefficient for the lower energy allowed a higher temperature rise during irradiation, leading to a higher tumor-killing efficacy. The PT conversion efficiency was higher for the 1064 nm irradiation as well, determined from the formula in Eq. (21) [285]. Another composition is one where Au was composited with CuS, allowing excitation of the LSPR in both and possibly bettering the PTT metrics. Indeed, the absorbance intensity of the composite when in the same concentration as when either of the materials were present alone was quantitatively higher [286]. Previous research has shown that the field enhancement “tails” of the LSPR can extend to energies much beyond the peak LSPR energy, resulting in much higher absorbance intensities due to coupling between this tail (of the Au’s LSPR) and the LSPR of the CuS. Specifically, the interparticle dipole coupling can create hotspots [287] in the enhanced fields, which can result in the observed effects. The killing of tumor cells was achieved at much lower laser powers in the Au/CuS study, the reason being attributed again due to the coupling between Au and CuS. This is likely (given that the excitation wavelength for the therapy was 980nm, much lower than the 530nm required for exciting the LSPR of Au), yet it is probable that other reasons are at play here, such as possibly a higher surface area for heat transfer of the composited particles which improved the killing efficacy. The NIR LSPR of CuS alone has also been explored for combined PT and positron emission tomography (PET) imaging by incorporating the radio isotope of ⁶⁴Cu into CuS which was synthesized in the nanodot morphology with an effective hydrodynamic diameter of less than 6 nm (controlled through the molecular weight and concentration of PVP). Irradiation at 808 nm resulted in a temperature rise to 50 °C, which along with the quick renal clearance (95% removed within 24 h) resulted in a composition that has a rare potential for dual modes of imaging and therapy of cancer [288]. Another commonly used radioisotope of ¹³¹I was also investigated for the same purposes, also proving to be effective in NIR absorption-based PTT and imaging when incorporated with an effective diameter of 20 nm, albeit with a protective cap of PEG [289].

$$\eta = \frac{hS\left({T}_{max}-{T}_{surr}\right)-{Q}_{S}}{I(1- {10}^{-A}\lambda )}$$

(21)

where η is the PT conversion efficiency, I is the laser power, A is a constant, λ is the wavelength of the laser, h is the heat transfer coefficient, S is the surface area of the well, T is the temperature of the solution, and Qs is the heat associated with the light absorbance of the well and solvent.

Insertion of different elements into the chalcogenide lattice leads to optical properties not possible with the chalcogenide lattice alone. Copper iron sulfide is a promising variant of CuS due to its extremely diverse LSPR bands. Plasmons absorbing from 486 to 1200 nm have been reported in this compound by a simple increase in the ratio of Cu: Fe from 1:1 to 10:1. High-resolution XPS studies revealed that the proportions of the cations and anions at the surface, the nature of the valence state of the surface atoms, and the nature and concentrations of the capping ligands decided the intensity and shift in LSPR position, determined directly by the Cu/Fe/S ratios used during synthesis [290]. The origins of the LSPR have hence not been explicitly dealt with in this study, and even a previous review on this composition only briefly ascribes that the possibility of the presence of side phases/Cu deficiencies is the reason for the plasmonic effect [291]. Clarifications in this regard were presented in a more recent work, wherein a changeover from an electron-based LSPR (at approx. 500 nm) to a hole-based NIR LSPR (at approx. 1200 nm) was observed wherein iron sulfates were produced during this transition, creating the hole carriers [292]. Yet another ternary composition is copper indium sulfide, whose Cu deficiency (resulting in Cu_xIn_yS₂) led to sub band-gap LSPR resonances in the NIR at around 1200 nm. The close energy spacing between the LSPR frequency and the excitonic band-gap (at 800 nm) makes this material promising for applications looking for light harvesting using both excitations such as light harvesting and catalysis [293]. Such ternary materials are often explored only for their band-gap tuning and seldom studied for their non-stoichiometric compositions [294]. Apart from the brief discussions on the possible elements that can be incorporated in the CuS lattice, a more comprehensive summary of others is presented in Table 3. The hitherto discussed possibilities of tuning the light absorption of the Cu_xS family by stoichiometry manipulation have been extended by an interesting investigation on this manipulation. The conversion between covellite and stoichiometric (non-plasmonic but fluorescent) chalcocite has been reported by electrochemical mediation. Specifically, the intercalation of Li ions into the chalcogen lattice led to this conversion, while Cu⁺ intercalation led to a permanent conversion to chalcocite. The prospects of these results is attractive for applications such as optical switching in the NIR and those involving fluorescent imaging such as sensing and other luminescence based applications [295].

The closing remarks of the CuS family of plasmonic materials are ironic with a review of a study attributing plasmonic absorption in stoichiometric CuS, which as a pristine material was not thought to not possess vacancy-originated hole plasmons. This study employed the discrete dipole approximation and the Drude-Sommerfeld model to account for the observed experimental spectra. The excellent fit between model and experiment was proof that the free holes necessary to exhibit the Drude-like behavior were constituents of the pristine lattice itself, with densities of up to 10²² cm⁻³. Additional confirmations of the inherent resistance to stoichiometric changes of CuS were also reported as additional proofs for this phenomenon. The nano disk morphologies investigated in this work exhibited two resonances in the NIR corresponding to one in-plane and another out-of-plane resonance [273]. Few studies exist on the LSPR origins of covellite such as this that attribute the LSPR nature to inherent lattice free-holes. It can be predicted then that these free carriers can make CuS reactive. Similar to how an oxide layer of controlled thickness can help in stabilization, the incorporation of Pd into the CuS structure as a CuPdS shell tuned the LSPR along with increasing the stability. A limitation, however, was the increased broadening with the increase in Pd content, due to a reduced particle size (and a resulting higher surface scattering) [296]. It is hence evident that the CuS family although exhaustively investigated is still nascent in the aspects of studying and manipulating the origins of the LSPRs in it.

Cu Chalcogenide Selenides

Non-stoichiometric Cu₂Se is another interesting prospect in the plasmonic copper chalcogenide group. The absorptions in this material have been in multiple instances attributed to impurity level transitions similar to the sulfide family and have since been corroborated well with LSPR excitations [297,298,299]. Mechanisms of achieving non-stoichiometry involve commonly the use of oxidizing agents or atmospheres and reversal through reducing agents [250, 268]. Unusual techniques such as ligand exchange which allow for tuning over 200 nm of the LSPR peak position [300, 301] are among the different modalities of tuning. Remarkable tuning of the LSPR can be possible, as shown in Table 2.

Table 2 Plasmonic copper chalcogenide selenides and the mechanisms of tuning their morphology and plasmonic properties

Electrochemical redox reactions can give useful insights into the mechanism of LSPR tuning in Cu₂Se, wherein in situ tuning and LSPR observation are possible [312]. Influences of the potential sweeping voltages (magnitude/sweep rate), electrolyte concentrations/compositions, hysteresis in material composition between redox cycles, and dispersion of the plasmonic NPs in a secondary medium such as a membrane (which itself will influence LSPR through particle aggregation/charge transfer) are a few of the multiple factors that influence the LSPR of the resulting NPs. XPS measurements confirmed the contributions to LSPR of Cu vacancies rather than oxidation state changes, as no peaks corresponding to divalent Cu were observed.

Non-stoichiometric Cu_2-xSe has also been explored for theranostics, such post encapsulation with a bio-compatible polymer that prevents cytotoxicity while still maintaining the particle sizes (10–50 nm) for longer retention within the body. The temperatures attained matched those when AuNRs were used, which makes them more suitable due to the better sizes (AuNRs were larger in order to be able to absorb at 808 nm) and the fact that the structure-directing agent of CTAB used for AuNR synthesis is cytotoxic [313]. The multifunctional role of plasmonic Cu_2-xSe nanoparticles (with sizes in the sub-5nm range) wherein the purposes of imaging (PTT), Single Photon Emission Computed Tomography (SPECT) and Computed Tomography (CT) could be possible was fulfilled by these nanoparticles. With regards to the plasmonic role, the temperature increases achieved as a result of the PT effect could be precisely controlled between 25 °C and 75 °C by changing the concentration of the NPs keeping the irradiation power constant [314]. A similar objective of sensitive photoacoustic imaging (pulsed optical excitation leading to ultrasonic wave emission due to thermoelastic lattice expansions) has been reported with Cu_2-xSe NPs for imaging lymph nodes, the usual paths of cancer progression. Highly sensitive imaging of the lymph nodes could be done with the concentrations of the phospholipid encapsulated NPs down to 1.7 pmol/ml while excited at 900 nm [315]. NPs are conventionally promoted as nano-sized carriers of drugs, and a composition of Cu_2-xSe encapsulated by a SiO₂ shell has been used in this regard both as a PT and chemotherapeutic nanomaterial. The cancer drug doxorubicin was loaded onto the shells along with a PEG cap that added the property of hydrophilicity to the complex, and the drug released by the heat generated on NIR radiation of the NP raised the temperature to 45.3 °C [316]. In place of an encapsulated shell, heterogeneous NPs (heterodimers) composed of Au and Cu_2-xSe have been reported for the combined possibilities of dark-field as well as photoacoustic imaging. The dark-field imaging was feasible due to the scattering of light by the AuNPs at 566 nm (red-shifted from the expected peak at 520 nm due to hole transfer from the Cu_2-xSe), enabling highly localized imaging of the infected lymph node. The PTT potential of the approx. 10 nm diameter NPs were not investigated, however, which should have been a possibility considering other research that has done very similar explorations into this aspect [317].

The plurality of the CuS/Se family in terms of applicability has been investigated with a rare perspective of their suitability for non-linear absorption. Specifically, the enhancements to the two-photon emission intensities from these materials catalyzed by LSPR have been studied, wherein the LSPR provides a virtual state for excitation. The mechanism of enhancement of this emission has been shown in Fig. 13 [318]. The explicit role of LSPR in enhancing this intensity was confirmed by tuning its absorption close to the two-photon absorption edge (by substituting Se for S at an atomic percent of 2.6), which was at 950 nm. A significant enhancement was observed for LSPR at this energy, compared to 1145 nm for stoichiometric CuS [318].

Fig. 13

Absorbance profile for plasmon-assisted two-photon-absorbing inorganic nanoparticles

Similar to CuS, insertions of elements in the Se-based Cu chalcogens have also been studied. Such ternary compounds exhibit variations in their LSPR based on their size and shape as can be intuited, but an additional modality of compositional tuning was achieved with varying the Se precursors with the same Cu and Sn precursors (leading to differing LSPRs due to differing free carrier densities). Specifically, the reason for this was the differing abilities of the different Se precursors to incorporate Sn into the lattice, which ultimately controlled the Cu:Sn ratio and hence the LSPR position and width, apart from the differences in morphology [319].

Plasmonic Cu Chalcogenide Sulfide Selenides

Plasmonic chalcogenide compositions of both S and Se have been reported in the interests of discovering the tunability possible in the optical properties for applications such as diodes. In this regard, compositions of Cu_2-x (S_ySe_1-y) were synthesized, and observations of the different stoichiometries indicated that the plasmonic behavior was predominantly controlled by the Cu deficiency and, to some extent, the phase of the materials (cubic vs hexagonal), and not the S/Se stoichiometry. The particle sizes were not found to be a governing parameter, as they were in regimes where the quasistatic approximation (predominantly dipolar contributions only and no multipolar contributions) holds as well as where surface dephasing was not an issue as well [320]. Another study of these compositions revealed an observation of variation in the direct band gap with S/Se ratios, although the changes to the free carrier absorbance were not investigated here. The purely cubic nanowire morphologies were synthesized maintaining the Cu ratios constant and tuning the S/Se ratio [321]. The tunability of the phase and the free carrier absorbance with the S/Se ratio was, however, reported through the synthesis of these compositions with different precursors and processing agents [322]. Achieving monodisperse particles as well as a broadly tunable (from 1350 to 1600 nm) LSPR of this composition was possible while using a specific complex capped precursor, while changing the specific complexes did not result in the same level of tuning [323]. One of the more comprehensive investigations of the insertion of Se into a CuS chalcogen through anion exchange corroborated the influence on the effective mass of free carriers, band gap, and the LSPR peak position with this insertion. The trends of decreased electron density, effective mass, and increased bulk plasma energy as well as LSPR were all attributed to the insertion of the larger Se atom which led to a more diffused setting of the lattice. However, experimentally little influence (a change of only 0.1e V in LSPR energy) was found on the insertion of Se, as unaccounted for effects such as the surface capping ligand and a decrease in free carrier density due to an increase in Cu content during synthesis were attributed to be more influential in reality [324]. It could be concluded then (at least from the reports from the majority of studies on Cu/S/Se) that the Cu deficiency and phase predominantly influence the free carrier absorbance, whereas the S/Se ratios influence the direct band gaps in these materials.

Plasmonic Cu Chalcogenide Tellurides

The enhanced scattering properties of either molecules in the proximity of LSPR active nanomaterials or of the LSPR nanomaterials themselves have been well established, including in reports in this review. In consideration of one of the mechanisms of this enhanced scattering, SERS, the distance between the molecule under study and the LSPR nanoparticles is of considerable importance for the enhancements. The ideal molecule will have moieties that can directly bind with the LSPR nanoparticle. With this purpose, the poster children of SERS (Au and Ag) have electronegativities which prevent their binding for example with –OH moieties which are a predominant part of many SERS active molecules. An alternate LSPR composition in this regard is CuTe, wherein the electropositivity of Te allows probing of molecules with –OH groups, enabling and extending the classes of molecules that can be investigated. The LSPR of CuTe being in the NIR is an added benefit in terms of the coupling between the LSPR mode and the Raman modes of molecules. Although comparisons of SERS enhancements between Au, Ag, and CuTe are not possible in most cases due to the inability to match the proximity between the molecule and the plasmonic material, considerable enhancements (approx. 10⁶) were observed with CuTe nanocubes and found to be the highest in comparison to nanoplates and nanorods of CuTe [325]. Interestingly, pump-probe spectroscopic measurements concluded that the time constants for LSP thermalization are lesser for Cu_2-xTe than Cu_2-xS and Cu_2-xSe by a factor of 3 and 2, respectively, potentially indicating their applicability for PT purposes as well [326]. It remains an observation that the plasmonic properties of CuTe through experiments and for different applications remain largely unexplored, though multiple reports on its synthesis exist [327]. Insights into field enhancements [328] are often used to validate the potential for a material for a certain application, and from one such investigation, Cu_2-xTe having morphologies of spheres, rods, and tetrapods were studied. Revealing insights on the distribution and excitation of the LSPR for these shapes indicated that the enhancements were found to be strongest for elongated shapes (rods and tetrapods) compared to spheres. However, the low carrier density (compared to Au for example) and possibly localized (and not collective) oscillations resulted in a lower enhancement magnitude of the field compared to similar morphologies of Au. The localized nature of carriers in Cu_2-xTe is also reported in Ab initio structure calculations, wherein vacancy channels of 0.3 × 0.7 nm were revealed to be alternatingly formed along the {001} layers connected by low-density areas between them as shown in Fig. 14 [329].

Fig. 14

Cu_1.5 Te structure determined and refined by electron diffraction tomography (EDT). a View along [100] and [100] highlighted in blue. b View along [010]. HAADF-STEM images are highlighted in blue. c View along [001]. Te represented by green, Cu atoms by red (reprinted with permission from [329])

The lack of a database for the material constants of this chalcogenide is also an issue here [330], as it is quite possible that the enhancements are not reflective of the actual numbers due to the difference in the LSPR peak positions of the calculations vs the synthesized nanoparticle, and hence, the field enhancement magnitudes are quite possible to be revised as research progresses on these materials. This is indeed made clear from a previous report [217], wherein field enhancements comparable to many if not the best coinage metals have been observed.

Combining the optoelectronic properties of semiconductors and metals can be of immense benefit, as both excitonic and plasmonic properties can be exploited. To this end, compositions of CdTe-Cu_2-xTe have been investigated. An interesting behavior of the Auger-mediated recombination between excitons in CdTe with the free carrier holes in Cu_2-xTe was reported, resulting in a quenching of the PL of the CdTe subunit although the investigators were careful not to rule out the changes to PL as due to the different surfaces and proximities for the CdTe vs CdTeCu_2-xTe compositions. Another observation was the unexpectedly low shifts in resonance frequency between the two materials, which was ascribed to the fact that the free carriers were localized more in comparison with the conventional free electron-based resonance exhibiting materials wherein the delocalization results in a wider influence on the plasmon resonance. This degree of localization was alluded to as the reason that the environment had a lesser impact on the resonance of the composite in terms of the shift despite the markedly higher refractive index. The carrier dynamics were found to be significantly altered for the composite than pristine CdTe material, ascribed also due to the faster Auger route of exciton relaxation. Such interesting combinations can be used for manipulating the decay dynamics and other interactions between these absorptions, which have been proven to lead to new phenomena such as Fano resonances [331].

The reports reviewed so far have been chosen to reflect the differences possible in terms of observations and applications of the predominant plasmonic chalcogenides of Cu. We present a comprehensive summary on the different chalcogenides of Cu used for plasmonic investigations in Table 3. Studies on the optical metrics of such Cu deficient chalcogenides aimed at revealing the absorption and scattering cross-sections, dependence of them on particle morphologies and the dielectric environment, and changes to them based on the myriad of possibilities present in terms of lattice alterations are vital to take these group of materials forward for such applications. Research such as reviewed here reveal the fundamentally interesting prospects of non-stoichiometric plasmonic chalcogenides, with regards to tuning their optoelectronic behavior in accordance with LSPRs as well as the coupling between them and excitonic transitions. This coupling for example leads to changes in the excitonic band-gap as well and is often reconciled by pertinent theories such as the Moss-Burstein effect, AC Stark effect, and the Purcell effect. Phenomena such as enhanced two-photon absorption, dynamical Stark shifts, up-conversion, and plasmon generation by quantized exciton generation are a few examples of potentially rewarding investigations in such plasmonic systems [332].

Table 3 Copper chalcogenides and mechanisms of tuning morphology and plasmonic properties

Plasmonics of Composited/Ligand Stabilized Cu Nanoparticles

We discuss here the design, synthesis, and study of materials and devices based on combining plasmonic Cu with various moieties as ligands, stabilizers, and generically as additives. The concept of MEF has been introduced already, and the use of metallic alloys composed in part of Cu has been investigated for their enhancements to the fluorescence of suitable fluorescent molecules. In this regard, the effect of adding Cu to the most commonly used AgNPs for fluorescent enhancement has been investigated. The result was that the CuNPs quenched the fluorescence when present in increasing mole fractions, as new channels of non-radiative decay of the absorbed photons were created in Cu. The overlap of the emission peak of the tested fluorophore (Cy3) with the inter-band transitions in Cu can be intuited to be a problem as well, as it will further quench the emission intensities. However, engineering the particle sizes to tune the scattering vs absorption (i.e., radiative vs non-radiative channels) percentages can help, as it was found that alloy particles of 60 nm diameter were found to cause MEF when present as spheres and in a dielectric of polyvinyl alcohol [370].

The composite of graphene/Cu has been projected to be one of the better prospects for photonic applications, due to reasons like a strong confinement of the plasmon at the interface, an effective protection against oxidation, and minimal interference with the plasmonic field enhancements due to the monoatomic thicknesses possible. As grown Cu/graphene can be advantageous (albeit having a higher defect density as a result of the lattice mismatch during growth) compared to transferred graphene grown in a separate CVD process wherein a better level of intimate electrical contact and adhesion is possible with the former. To overcome this limitation, CuNPs were formed post deposition of a continuous film simultaneously during the deposition and growth of Graphene with methane as the carbon source. The resulting capped composites exhibited a tenfold enhancement in the fluorescent intensity of a DCM dye due to strong coupling and energy transfer between the plasmonic and dipolar modes [371]. Graphene-protected Cu was also studied as a fluorescence-enhancing combination for a single crystalline material. BP3T modified with this combination exhibited a marked increase in PL intensity when compared with either BP3T/Cu or BP3T alone. The contribution from graphene was not explicitly addressed (apart from improvements to oxidation stability), while the enhancements to the intensity were attributed due to the coupling of the LSPR of Cu to the emission from the dye and a reduction as a result of the emission lifetimes. Directional emission studies measuring the polarization of the emitted light could have shed more light on the causal reasons for the enhancements [372]. An interesting aspect to be mindful of while designing graphene-protected Cu is the presence, concentration, and nature of defects. Comparing the defects present in sheets vs in graphene powders led to a change in the intimacy in contact between them. The result was a lower H₂ evolution (through H⁺ reduction on graphene defect sites) when Cu alone was used and when graphene with lower concentrations of defects was used. The mechanism of hot carrier generation through plasmon decay and the crucial subsequent step of their separation (electrons to graphene and holes to lactic acid) was proposed as the causal mechanism for the observations. Hence, it can be stated that graphene with defects can be beneficial too, as evidenced by this example [373]. Graphene has hence become a strong associative combination with Cu for a myriad of purposes and enhanced phenomena.

Mimicking natural processes has been an ongoing effort for a diverse range of purposes [374]. One such process which when mimicked can give efficient results is photosynthesis, in which utilization of CO₂ forms a part. This begins with conversion of CO₂ into for example the CO₂ radical anion or CO, which can initiate multiple reactions such as fuel production from biowaste. For the catalytic reduction of CO₂, plasmonic materials are of benefit due to the generation of hot carriers during thermalization. If the hot electrons and hot holes can be separated, reactions can be catalyzed [375]. In this research on hot carrier generation and separation, the contributions to the hole population (when for example d-holes are generated) by plasmonic decay are crucial [376], as the energies of these holes critically dictate their lifetimes and hence probabilities to participate in reactions. This dependence of the d-hole energy on its lifetime is directly tied to the DOS which for Cu varies steeply with energy as shown in Fig. 15 [376] compared to for example Au, as the hybridization between the d and sp bands is stronger for the latter leading to a broader DOS and hence a lower dependency on energy. The result is that for d-holes created at levels at increasing distances from the Fermi energy, the lifetimes increasingly reduce due to this nature of the DOS in Cu [377].

Fig. 15

Schematic diagram of interband emission involving d-holes created by 488, 532, and 633 nm excitations on Au vs Cu

Meticulous calculations of the lifetimes confirm this trend, with 250 fs observed for hot electrons excited to energies of 0.1 eV above the Fermi level whereas for those excited to 2 eV a drastically reduced lifetime of 20 fs [378]. In this field of hot carrier separation catalysis, the natural tendency of Au NPs to be more inert than Cu has made it and Ag the main proponents for this field of research, and yet Cu has been studied extensively for its efficacy in catalyzing CO₂ reduction, albeit not very often on the plasmonic angle but through electrochemical reduction catalyzed by it [379]. Meanwhile, a comprehensive and careful summary on plasmonic catalysts in general that exclusively treat CO₂ reduction alone is available [380]. The binding energy of CO on Cu is the critical factor in determining the investigated materials’ selectivity and catalytic activity [381] with Cu occupying the middle ground in terms of the binding energy and metals like Ag, Au, and those of Pt, Ni, Ru, and Rh occupying the low and high binding energy regions, respectively. There are studies, however, that have investigated this enhancement in CO₂ reduction using alloys of Au-Cu on TiO₂, yet do not ascribe hot carrier separation-induced reasons for the same [382]. Other reports for example have investigated the synergistic effects of a Cu/Pt/TiO₂ composite as well as an Au-Cu alloyed SrTiO₃/TiO₂ composite and report an increase in CO₂ reduction based on the increased photo-generated carriers from TiO₂ [383, 384] and only hold Cu to be responsible for presenting sites for adsorption/desorption with no mention of the plasmonic enhancements from Cu itself. More published research points to the same, wherein the plasmonic contribution by Cu was not investigated (probably due to the nature of the surface Cu species to be oxidized), and its role was as a recombination preventer/adsorption enhancer [385]. A brief mention of the enhanced electromagnetic fields of the Cu plasmons being a contributor to the enhancements was made in this study [386].

The Au-Cu Alloy System

Au being the other poster child of metal photoluminescence has been alloyed with Cu in an extensive array of reports. Careful studies carried out through varying the lattice proportions of Cu, particle sizes, binding ligands, and excitation as well as emission wavelengths revealed that the NIR PL emission from the alloyed NPs arose from transitions within the Au-thiolate (the most often used cap for Au clusters) surface complex. The Kohn–Sham potentials used to reveal the orbitals and the DOS corresponding to this system confirmed this postulation along with excitation spectra, with little variation in the PL position with size. Increasing amounts of Cu resulted in an increasingly redder emission, with atomic percents from 0 to 80% shifting the PL peaks from 947 to 1067 nm [387]. The brightness of these alloyed NPs was also brighter than the best NIR-emitting lanthanide probes, indicating their deployment readiness for various luminescence-based applications.

Furthering such investigations, Au-Cu alloys have been proposed to increase the conversion of nitroaromatics to anilines predominantly through the reason of Cu providing a stronger site for adsorption than Au due to its slightly electropositive nature [388] when present as an alloy. The Au-Cu alloy system has been a strong contender in this field of plasmon-enhanced catalysis due to tunable absorbance and nature of surface sites [389], with studies that report on their uses for example in enhancing rates of conversion of 4-nitrophenol to 4-aminophenol with alloyed triangular prisms [390]. A summary of this system and the reactions studied with the investigation of its plasmonic advantages is provided in Table 4.

Table 4 A summary of Au-Cu alloy system to enhance the plasmonic resonances and the mechanisms of tuning their morphology

The first observation of the interactions between the constituents in a plasmonic alloy is the change in absorption. Among the myriad investigations that often confirm a bathochromic shift with increasing Cu atomic fractions, ordered Au-Cu alloys have also shown a hypsochromic shift, attributed to the ordered creation of the alloys in contrast to a random formation by modifying the synthesis process [407]. Compared to other alloy systems that will be discussed next, computational findings suggest that Au-Cu can be the most favorable composition when looking for a homogeneous mixing of both atoms instead of the atoms becoming segregated. The thermodynamic driving forces calculating the energy of formation predict a decrease in formation energy of Au-Cu compared to Au alone for appropriate proportions of Cu insertion, confirming this conclusion. The close overlap between the Au-5d and the Cu-3d orbitals allowed charge transfer to the more electronegative Au from Cu, and though the simulated clusters consisted only of 100–200 atoms, they exhibited a collective “plasmon-like” electron oscillation due to the close overlap [408].

The presence of Au as a compositing element to Cu rather than an alloying element has also been studied. The resulting coupled plasmon being termed as a charge transfer plasmon (CTP), leads to interesting effects such as a drastically red-shifted resonance position (up to 1200 nm) compared to both Au and Cu. The nature of the contact line between Au and Cu proved to be a deciding factor of the resonance position and width, with adsorption of DNA on the preliminarily synthesized Au seeds also controlling this contact line [409]. Such structures when studied in a more comprehensive manner through multiple investigations can be promising especially in terms of applications concerned with surficial properties, such as sensing and catalysis.

The Cu-Ag Alloy System

Another alloy system of Cu is with its coinage peer of Ag. Interestingly, the Ag-Cu system has been reported to form alloys as well as bimodal compositions, wherein the latter can be present as a core–shell arrangement or as segregated phases of Cu and Ag with separate LSPR peaks for each. Both compositions have been reported with an almost equal occurrence. The bimodal composition being investigated for its plasmonic properties is summarized in Table 5.

Table 5 Plasmonic properties of the bimodal composition and the mechanisms of tuning their morphology and the prospective applications

Studies have reported the possibility of controlling the formation of products (azo vs azoxy) through the use of Cu-graphene vs Cu-Ag-ZrO₂. The former resulted in the formation of both azoxybenzene and azobenzene from nitrobenzene, while the latter resulted in the formation of azoxybenzene from nitrobenzene. For the Ag/Cu/ZrO₂ study, multiple investigations were done to rule out/confirm this catalytic activity, ranging from XPS to confirm the metallic state of Cu in the alloy vs when present alone, elemental mapping from TEM to confirm the presence of both Ag and Cu in the lattice, and catalytic activity measurements to confirm that only the alloy enhanced the conversion in an optimal amount compared to its monometallic counterparts, and that only the alloy resulted in a selective conversion to azoxybenzene instead of azobenzene. TEM analyses were concluded to reveal the (111) plane of the alloy’s lattice as a governing observation, albeit confirmation through other studies/with calculations could have revealed insights into the DOS distributions for example. The plasmonic peak was observed at approx. 408 nm for the alloy, implying a closer match to that of Ag than Cu. This could be reconciled from the proportions of the Ag: Cu for the alloy, which at values of Cu higher than 4:1 led to a considerably oxidized surface and a loss of catalytic activity. A higher conversion was observed for excitation wavelengths close to 400 nm, implying a plasmon-induced enhancement to the conversion as well as the observation of a lowered activation energy for the irradiated samples confirming this postulation. The study with graphene as a stabilizing support (for Cu in the metallic state) reports the plasmon peak and the associated enhancements to the conversion due to an efficient cleavage of the N–O bond. The conversion efficiency was found to markedly depend on temperature due to the different activation energies and the Bose–Einstein electron distributions [414]. Considering that visible light irradiation has been employed in these studies, it is quite probable that contributions to the effective free carrier densities in the alloys from Cu play an active role to the catalytic rates as well, although the contribution is probably detrimental considering the easy tendency for oxidation of Cu. It is hence a compromise probably that the addition of Cu provides considerable tunability in the LSPR position, while limiting the stability of the pristine metal that it alloys with to a very short period of time (usually days). Reports confirm this observation for the Ag-Cu system as well, wherein tuning from ~ 400 to ~ 600 nm is possible with different mole fractions of Cu albeit with accompanying stability limitations [381, 415]. A novel multi-compositional nanorod comprising of sodium yttrium fluoride (NaYF₄), TiO₂, and a Ag-Cu alloy has been investigated for two-photon up conversion-mediated photocatalytic applications with the aim of generating the highest currents on photoexcitation. The specific functions of upconversion in NIR by NaYF₄ and injection of the up converted photon into TiO₂ for carrier generation, of TiO₂ for carrier generation, and of the Ag-Cu alloy for preventing carrier recombination and to aid in photocarrier generation due to the overlap of the LSPR with the TiO₂ band-gap were elucidated in this study. The role of Cu however, was attributed only as reducing Ag utilization and hence costs, as substantial amounts of it will shift the LSPR away from the absorption edge of TiO₂ and reduce the photocurrent amplitude [416]. An analogous investigation with the composition of Ag-Cu-decorated SiO₂ on which layers of S-doped carbon nitride (CN) were wrapped on has been reported as an efficient photocatalyst. In this case, however, the generation of hot carriers and their transfer to CN (preventing recombination) which catalyzed the production of H₂ from water was the mechanism of photocatalysis. The molar ratios of 1:3 of Ag: Cu led to the highest H₂ generation rates among the compositions investigated and were optimal to prevent recombination (due to close contact) between the alloy and the CN wrapper. The comparatively lesser reduction in PL intensity for this ratio compared to others confirmed the attribution of this ratio as optimal (due to the absence of hot carrier recombination which would have reduced H₂ production) [417].

Other Plasmonic Cu Composites

The importance of the separation distance between the plasmonic material and the scattering center such as a fluorophore was investigated in a study employing a hierarchical structure of CuNWs and MoS_x dispersed in a Cu foam. Marked differences in PL intensities revealed the differences in recombination rates for the compositions involving MoS_x blended into CuNWs and Cu foam vs those where CuNWs were first blended with MoS_x post which they were dispersed in the foam support, making the former much better for photocatalytic H₂ evolution [418]. Transmetalation of ligands can also be catalyzed by LSPR, and from the one report on this specific reaction, the irradiation wavelength was reported to be crucial. The switch between homo-coupling to cross-coupling when irradiated with different LSPR active wavelengths (< 500 nm vs > 500 nm) using Cu-Pd particles was attributed to be the reason for the different reaction intermediates that were formed under these conditions. Interestingly, the transmetalation was not observed when using pure Cu or when Cu was combined as a composite with other molecules such as Pt/Al₂O₃. A brief attribution to plasmonically generated hot electrons in Cu was given for the differences in selectivity, and the presence of Pd was critical for better adsorption kinetics, Cu-Pd coupling, or other yet undiscovered factors [419]. Considering the plasmonic metal of Pt, an interesting investigation wherein hybrid propagating modes that arise due to coupling between a specific periodic arrangement of nanostructures and the plasmon modes has been reported in the bimetallic system of Pt-Cu. Galvanic displacement of lithographically patterned Cu squares by Pt led to this bimetallic configuration, and the absorption signatures were dominated by Cu (interband transitions at 510 nm, dipolar plasmon at 675 nm) and a hybrid surface lattice resonance (HSLR) at 865 nm which was invoked due to the coupling. The photocatalytic enhancements were not found to be influenced by the Pt plasmon. Among the different experimental conditions tested to quantify the catalytic activity such as in dark conditions, with and without Pt, and with different shell thicknesses of Pt and size/spacing of Cu, the strongest enhancements arose in the presence of the HSLR, attributed to the intense electromagnetic fields at the boundary between the Cu/Pt nanoparticle and the ITO substrate [383].

The higher energies of plasmonically generated hot electrons compared to those generated by decay of interband transitioning ones have been predicted and confirmed, making the latter process more efficient than the direct inter-band excitation [420]. This stems from the fact that LSPR being confined close to the surface, only electrons near the Fermi level will be influenced by LSPR energy transfer, leading to a higher energy as they transition to the excited state compared with the electrons that are excited from lower energies below the Fermi level. Hence, exploiting plasmons for creating hot carriers is more advantageous than generating them through inter-band transitions. Towards further investigations of this plasmonic angle presented by Cu, p-NiO has been used to effectively accept hot holes, thereby segregating the hot electrons in the Cu. It is important to have a semiconductor that is transparent in the range of plasmon absorption (example: a band-gap of 3.7 eV for NiO) such that the semiconductor’s valence band (example: − 5.4 eV in vacuum for NiO) lies below that of the metal (− 4.5 eV in vacuum for Cu) for hole extraction. The enhanced rates of production of both CO and formate (HCOO⁻) on visible light irradiation confirmed the role of LSPR-generated hot electrons in Cu, thought to form these products from the creation of the anion radical first. Importantly, the competing mechanism of H₂ evolution which always occurs along with CO₂ reduction was suppressed. The hypothesis that this is the case because of the increased dissociation of H₂ that does evolve on the Cu surface but dissociates through the DIET (desorption-induced electron transitions) mechanism has been put forward to explain the reason for this selectivity [421]. The DIET mechanism has also been proposed to be a reason for the reduced activation barrier for desorption of reaction products in the reaction of NH₃ to produce H₂ on a Ru-Cu alloy. The reductions in the activation barriers of all the reaction steps beginning with the scission of the first N–H bond to the final desorption of the products were predicted to be influenced by the charge transfer of hot electrons to the adsorbate orbitals, with enhancements to the reaction rates ranging from ~ 20 to ~ 177 compared to when only Cu or only Ru was present [422]. In another research studying epoxidation of alkenes including styrene, trans-stilbene, cis-stilbene, and norbornene, the required energy for the reaction was achievable only at energies higher than 500 nm and hence excluded plasmons from playing any role, except for cis-stilbene where the epoxidation was observed for the excitation wavelengths matching the LSPR absorption of the Cu-TiN composite nanomaterial [423].

Hot electrons, being of obvious interest in a myriad of applications involving reaction enhancements and facilitation, have been reported to be generated through the mechanism of plasmon relaxation even in sub-3 nm Cu nanoparticles. This is probably a controversial work reviewed here, with the ascribing of the inter-band absorptions observed (centered at 395 nm) to an LSPR absorption which is intra-band. The authors clearly ascribe the higher energy of this absorption due to band discretization but still attribute the same to LSPR. Regardless, a novel approach of controlling the distance between the sub-3 nm CuNPs and the support materials of TiO₂ to distances of approx. 0.42 nm was followed (with the help of multiple linker molecules), so that only hot electrons with the required energy (to cross the resulting Schottky barrier height of 0.88 eV) can transfer into the semiconductor and aid in the eventual evolution of H₂ [424].

A side note on the LSPR of small (< 3 nm) CuNPs is apt to be had here. Multiple studies such as this attribute a “shift” in LSPR to energies up to 6.2 eV (200 Nm) [65, 425] for the particle sizes observed, while the studies cited for these reports only mention a vanishing of LSPR at these sizes with no mention of such a marked shift in the LSPR energy, and clearly describe the impossibility of a continuous DOS (and hence an LSPR) at these sizes nearing 1 nm [426]. There are studies that report an absorbance peak analogous to plasmonic absorption for particles from 2 nm down to nearly 0.6 nm, with a marked decrease in contribution towards this absorbance by free carrier absorption (LSPR) as compared to bound carrier absorption (inter-band transitions). It stands then for Cu that for sizes less than 10 nm (the mean free path for Cu), the plasmon is highly damped, and that the LSPR contribution to absorption is substantially negligible at sizes less than 2 nm where energy band discretization into discrete energy levels occurs [427]. With all the investigations on plasmon-enhanced catalysis, commercialization presents fresh challenges in terms of cost, efficiency, precision, etc. Towards addressing these issues, amine-capped CuNWs were through a facile process of drop-casting dispersed on a fabric, enhancing porosity and dispersion simultaneously. Plasmon-enhanced PT temperatures of up to 250 °C were achieved, resulting in this configuration being a viable proposition for the catalytic azide-alkyne cycloaddition reaction. Furthering this investigation to include the reproducibility and optimization and affirmation of the role of morphology (as no distinct plasmonic peak was reported and yet was attribute to as a reason for the high absorption of incident energy), for example, can make such ligand-stabilized nanostructures marketable [428].

A survey of research reveals that PT therapy (PTT) is more effective when employing composites or alloys instead of pristine materials. This is primarily (from a plasmonic point of view) because of the ease of tailoring the dielectric constants (which determine the proportions of irradiated energy that is scattered vs that which is absorbed) and the ease of tuning the absorption energies. Both of these abilities are vital, when considering for example that the breadth of absorption energies of only the pristine materials has limitations. Cu has been investigated for PTT with this reasoning primarily as sulfides and selenides as discussed in this review. However, novel compositions have been attempted, with promising results. As an example, Cu-Ag₂S has been synthesized as an alloy with both Cu and Ag in the + 1 oxidation states, indicating a reduction in free electron density. This reduction made possible the peaking of the LSPR of this alloy in the infrared, making it apt for PTT. The reported PT conversion efficiencies (58.2%) were better than the CuS class (27.1%) when the same loadings were used. Careful measurements of the reasons for this, such as the dispersion, retention times, and the effective dielectric constants, can reveal the complete understanding of the material functioning [429]. Another example of an alloy composition with Cu is with Pd, with the Cu component being added for the sole purpose of aiding in autophagy [430] of the tumor cells. The LSPR properties of the composition were not investigated in detail apart from a brief mentioning of the sharp tips of the formed plasmonic tetrapod causing a temperature increase of up to 71 °C on irradiation with an 808 nm laser. Justification of the shape being the major causal factor for the improved PT heating was done by comparing the heating efficiency with spherical nanostructures wherein the latter resulted in much lower temperatures (approx. 45°) realized due to the less intense field accumulations [431]. Investigations of modifications to the lattice of this system can lead to more aspects of plasmonic manifestation such as the phase modifications, stabilities against chemical and thermal environments, causes of the observed morphologies, and the extinction cross-sections, as both these elements have explicit benefits in applications such as H₂ sensing and storage and thermal energy storage and transfer. Indeed, such an investigation has been done for the more common Au-Cu alloy system, wherein the five-fold twinned edges of the initially formed seeds were found to be the growth planes, leading eventually to a starfish-shaped nanostructure. The differences in particle sizes (ranging from 70 to 200 nm) were hypothesized as the reason for the observed plasmonic absorbances in the visible to the infrared. Notably, the atomic percent of only one particle size was quantified. We believe that compositional studies of each size could have revealed differences in atomic percent between Au and Cu, which could have added an additional variable that caused the observed LSPR characteristics. The 70 nm particles were deemed more suitable despite the poorer temperature rises achieved (due to less overlap between the 808 nm excitation and the NIR LSPR for this particle) as they could be retained for a longer duration in vivo post coating with a bio-compatible layer of PEG [402].

Optoelectronic materials, concerned with tuning the band gap for use as light emitters, are often explored with the CZTS/Se (Cu, Zn, Sn, S/Se) class of materials. The plasmonic behavior of these materials has been extensively reported including in the discussion in this review. One study projects their suitability for solution-processed photo-voltaic devices due to their easy reversible gelation ability [432]. The LSPR peak was centered at 1326 nm, with variations in position as well as intensity achieved by varying the Cu concentrations. An interesting study that observed the intermediate phases that led to the final composition of Cu₂ZnSnS₄ through SERS reported the formation of the plasmonic Cu_2-xS phase in the beginning stages, which gradually disappeared as the final composition materialized [433]. In the prevalent optoelectronic application of solar cell fabrication and improvement, CuNPs were found to increase the efficiency of a PEDOT: PSS-based perovskite solar cell by 115% due to the LSPR property. Specifically, time-resolved photoluminescence experiments revealed the shortening of the radiative recombination of photo-generated carriers due to the strong coupling of the plasmon to the host material as the reason for the efficiency enhancement. A contradiction in their conclusion must be noted here that the least average lifetime of emission was observed for a particular concentration (2 mg of Cu per ml of the composite solution), and yet this concentration had a higher lifetime for the plasmonic component and a lower lifetime for other less influential components governed by traps/defects, leading to the contradicting but inevitable postulation that the reduction in average lifetime does not always imply an enhancement to scattering. It can be understood from the results of the PL emission experiments [434] that the significantly increased intensity of PL for the 2 mg Cu sample could be the reason for the observation of the best efficiencies. Similar cursory conclusions of the enhancements to photocurrents being due to LSPR in solar cells based on Cu-PEDOT:PSS and Cu-TiO₂ have been made in other studies, with necessary investigations into the other aspects such as the influence of the interface/particle diameter and shape, coverage, and durability towards oxidation unaddressed [435,436,437]. An illustration of the reflectance of Cu as a function of wavelength overlapped with the solar spectrum shown in Fig. 16 [438] has been given to visualize the benefit of utilizing Cu for enhancing solar absorption for various applications including solar cell efficiency improvement. The selection, therefore, of the wavelength of absorption is crucial, wherein studies have reported enhanced magnitudes of photocurrent even in the proximity of Cu due to a plasmonic nanoantenna effect [439]. This effect, simply put, is the complete concentration of the incident energy into plasmon excitation, which led to a higher photocurrent due to more efficient carrier generation from plasmon-TCPP coupling. The promise of nanotechnology especially for plasmonics is the possibility of manipulation of a wave plasmon vs localized plasmon vs hybrid plasmons by a facile manipulation of morphology and/or the excitation configuration. Such manipulation has also been cursorily investigated for solar cell applications, specifically on the use of Cu nanowires for inducing a wave plasmon. Through the well-researched combination of Cu-TiO₂, the nanowire morphology was found to enhance the incident photon to electron (IPCE) conversion efficiency by 7.5% compared to devices with only CuNWs and by 24% compared to when only a dye was used for photocurrent generation without any CuNWs on visible light illumination. Combinatorial factors of the proximity of dye molecules to the TiO₂ and Cu, the Cu plasmon for generating hot carriers that entered the conduction channel in TiO₂, and an unsubstantiated “waveguide” property of a propagating SPP in the nanowire were attributed to be the reasons for the same [440].

Fig. 16

Reflectance of Cu as a function of wavelength compared with the solar spectral irradiance

Other Compounds of Cu Yet Unexplored/Emerging for Plasmonic Applications

Interesting chemistries of Cu have been synthesized, such as a mixed metal compound consisting of an antimony/Cu complex (((Sb₂(Ncy)₄)₂Cu₄) [441], CsCuTeS₃ [442], and CuIn_1-xGa_xS₂ [443] for which explicit treatments of the plasmonic behavior can be done. More examples exist, such as the mixed chalcogenide-based Cu_yFe₄Sn₁₂X₃₂ (X = S, Se) [444] and the CZTS/Se family of materials containing both S and Se such as Cu₂ZnSn(S_1–xSe_x)₄ [445,446,447,448] all of which were scrutinized for their band-gap values primarily, along with carrier masses and mobilities aimed at optoelectronic applications. Similar to the prospect of finding an intermediate/lattice component that is plasmonic in such materials like the Cu_2-x phase as discussed in this review, a plasmonic Cu_2-xSe_yS_1-y phase was formed as an intermediate in research that reported the formation of Cu₂ZnSn(S, Se)₄, wherein the plasmonic component was largely quenched while transitioning towards the final phase [432]. Other Janus-type nanoparticle compositions with distinct surficial properties have been synthesized, such as the composition of Cu-Cu₃P [449] whose LSPR aspects can be very unique once revealed through future investigations.

The property of LSPR hybridization with other propagating optical modes such as SPP has been explored for use as a directional glazing material, composed of periodically patterned Cu on a continuous film. By changing the position of the deposited substrate from one that has a continuous film of Cu on one side to one where the nanopatterned Cu is present, energy could either be reflected or be absorbed and converted into heat, making the substrate an ideal conformation as a coating for building windows to maintain indoor temperatures. Such metasurfaces can provide versatility in tuning the light extinction by controlling predominantly the spacing between the patterned features, leading to hybridized [450] modes with broadband plasmonic absorption.

Metal–metal alloy/composites comprised of Pt and Cu have been reported to be effective for PTT. However, the high absorbances in the NIR region had no signatures of typical LSPR absorption. Assessments of the lattice structure, LDOS calculations, and incident field enhancements of the dendritic structures of Pt/Cu achieved here could lead to these materials being extremely tunable agents for applications concerning light-to-heat conversion [451]. A highly multi-elemental composition of [Au₁₂Cu₁₃(Ph₃P)₁₀I₇] (SbF₆)₂ is among a common class of possibly plasmonic nanocomposites. The base configuration of such compositions is M_nL_m where n and m are the number of metal atoms and the protective surficial ligands [452]. Despite the effectiveness of these composites being proven for various applications such as non-degassed phosphorescence (i.e., in luminescence applications that require no de-gassing which typically causes PL quenching) and photocatalysis, the plasmonic aspects have, to the best of our knowledge, not been paid due investigations and hence merit analysis. The LSPR property cannot be completely reconciled in such clusters unless the particle diameter exceeds ~ 2 nm, corresponding to an atomic number of ~ 300 atoms for Cu. Yet, these intricate structures [452] with single atom level manipulations can likely exhibit LSPR characteristics by tailoring the number of atoms [453] and hence the free electron density to such appropriate levels, enabling their access to a myriad of applications concerning LSPR such as the capability to serve as hot carrier injectors and PT manipulators. Among the multiple bimodal/bimetallic alloy systems of Cu, the Cu-Ni system is a rarely reported candidate. With analogous observations of a red-shifted resonance with Cu content and particle size, this tuning of the effective dielectric constant can be of use similar to the other systems. Also, the fabrication procedure of the Cu-Ni bimetallic alloy through de-wetting of films from different substrates and the resulting differences in the particle diameter, distribution, and the diameter: height ratios can be a beneficial route for intricate exploitation of the plasmonic properties, extendable to other alloy systems as well [454].

Similar to chalcogenides, phosphides of Cu have been reported to be both plasmonic and semiconducting. To tune the LSPR in the mid-IR for these materials the nucleation controlling agent of Trioctylphosphine (TOP) was used, which showed a trend of inducing lesser Cu vacancies and hence a more-red shifted LSPR with its increase. This material too has promise due to the ability of tuning LSPR by tuning multiple parameters such as the concentrations of TOP, the surrounding material, and the shape and size of the achieved structure, although the latter was found to play a much lesser role in this investigation [455]. The plasmonic nature of this compound is being confirmed by observation of refractive index tuning as well as the inherent negative energy of formation of Cu vacancies, and additional confirmations came in from the observation of the p-type nature of the materials indicating a hole-based LSPR. A solid step towards comprehending the fundamental nature of the LSPR in this composition has been done through pump-probe spectroscopy which resulted in a non-linear and instantaneously bleached signal at approx. 1100 nm, indicative of LSPR nature [456]. The promise of these materials, however, has room for exploration, as their potential has only been mentioned in passing for applications such as photovoltaics and synthesis of cation-exchanged novel compositions. For photovoltaic applications aspects such as the illumination power and wavelength, play a critical role as confirmed by a study employing composited graphene/Cu_3-xP nanomaterials. The investigation revealed an increase in photocurrent when illuminated at the LSPR wavelength, but mechanisms causing this increase were not addressed although the generic cause of hot carrier generation and transfer can be intuited to have an influence [457]. Studies revealing the mechanisms for resonance and the locations of the Cu vacancies from EELS and atomic contrast STEM imaging revealed that the vacancies arise from the absence of Cu at lattice sites of type Cu1 and Cu2 rather than of types Cu3 and Cu4 [458], and that two modes of plasmonic oscillation are present. The first one corresponds to the dipolar mode while the second one to a non-polar breathing mode (also termed as a dark plasmonic mode as it cannot couple to light [459]) with radial oscillations that can be excited only with an electron beam. Such studies reveal the carrier concentrations, origins, and localizations of collective plasmons and hence are critical in establishing the potential of emerging materials such as phosphides for applications such as Q-switching and NLO [460].

With regards to other materials for which the LSPR has not been given credit but where it is probably a contributing factor to the observed improvements to properties, one composition is of CuNPs embedded in a CNT/pCN composite which was investigated for its improved photocatalytic CO₂ reduction [461]. A more detailed analysis of the plasmonic aspect of these materials is expected to unambiguously contribute to a variety of fields in ways which have been discussed here.

A very interesting proposition of LSPR is on generating low-frequency plasmons, specifically those that exist in the microwave frequencies. Although the plasma frequency of Cu lies in the visible, studies reviewed here have shown the dramatic changes possible in the peaking frequencies. As per Eq. (12), SPR occurs when the permittivity of the dielectric is negative compared to that of the metal. Remarkably, distribution of CuNPs on a three-dimensional matrix so that a threshold of percolation (usually a vol% of Cu, dictated by the particle size and the matrix’s morphology and composition) is exceeded led to this negative permittivity at microwave frequencies, with the induction of a delocalized electron oscillation. The creation of metallic networks was proposed to create this delocalized electron cloud of the individual NPs resulting in the collective excitation. Further studies on achieving/confirming/tuning of this low-frequency plasmon should reveal new opportunities for these materials such as in attenuating devices [462]. Such oscillations have been observed for multiple combinations of Cu, such as Cu/BaTiO₃ [463, 464], Cu/YIG, Cu/Ni-Zn ferrite [465, 466], and Cu/polyphenylene sulfide.

Studies on employing Cu as an enhancer of SSA which can tailor make the Cu architecture to be solely absorbing instead of absorbing and scattering is an example of research that can immensely benefit this area which is highly driven towards large-scale production. A summary of the different materials that have been used with Cu as deposited over layers for the purpose of enhancing SSA is presented in Table 6. Investigations of the interfaces in such architectures on various aspects such as the influences of Cu morphology, deposited film parameters, and other interfacial properties are essential to discover the relevancy of such compositions for plasmonic purposes.

Table 6 A summary of the different materials that have been used with Cu as over layers to enhance SSA

A wide array of materials has been investigated as partners of Cu for the specific purpose of enhancing absorption of solar energy. In all these studies, strictly cursorial treatments of the absorbances are present, with multiple questions on the plasmonic aspects unaddressed.

Finally, plasmons in non-stoichiometric and layered CuO have been found to be plausibly responsible for high-temperature superconductivity. The likelihood of the plasmons pairing with electrons for superconductivity was found to be high if the lattice was in a layered fashion, as the localized plasmon DOS of such a structure was found to be conducive for plasmon excitation and coupling with superconducting electron motion. It was, thus, concluded that along with other excitations such as phonons, excitons, and spin-fluctuations, plasmons also facilitated attractive electron–electron interactions and hence superconductivity [477].

Conclusions and Outlook on the Future of Cu Plasmonics

The multiple directions pursued in the research on plasmonic Cu have been attempted to be methodically covered in this review. Plasmonic Cu has come a long way since its discovery and ostracization soon after for plasmonic applications. The maturation of synthesis technologies according to the authors is an important and probably a spearheading reason for the continued expansion of Cu plasmonics, keeping in mind the changes to the concentrated field distributions that can be manipulated seemingly at will due to this maturation. Indeed, the coupling between far-fields of the LSPR of CuNPs with other excitations such as inter-band as well as hybrid excitations which has been discussed in this review underlines the possibilities in terms of material design [478]. The possibility of experimenting with combinations that are increasingly novel while standing on the shoulders of the foundational materials and technologies can be understood from the discussions on the various topics of this review. The control over morphology, composition, oxidation, dispersion in the case of nanoparticles, and analogous aspects of roughness, precision, uniformity, and purity while pondering over films have been enablers for this burgeoning of investigations. However, much space exists in terms of locking down plasmonic Cu as a material for which a complete understanding/application of its optical properties has been achieved.

To mention a few areas where improvements are possible, separating the contributions to enhanced catalytic performance due to plasmonic contributions can be better elucidated, as with eccentric but precise techniques such as scanning tunneling microscopy and EELS. Revealing these details can help in better exploiting and designing materials aimed at harvesting of unutilized energy (such as solar energy and waste heat), as well as designing better materials that can prevent carrier recombination and improve reaction rates. The authors opine for example, that the myriad of studies that report Cu (such as in the Au-Cu system) only to enhance the adsorption of reactant molecules with no part to play in plasmonic hot carrier generation (attributed to the Au component of the compound) though illuminated with visible light can be furthered to study the likely plasmonic contribution of Cu. Metal-enhanced fluorescence is one of the fascinating areas of research that can benefit by a margin from mechanistic insights into the phenomenon. With the importance of the proximity of the fluorophore to PC, their location within a material plays a critical role in determining the near vs far-field enhancements/quenching to MEF [479]. Studies such as pump-probe spectroscopy combined with simulations can be of use in this regard, as they have in the past in terms of revealing the sequence of events that probably lead to the enhancements. Bettering the illumination method to achieve a total irradiance of CuNPs and their conjugated fluoromers is another aspect of bettering fluorescence and scattering performance in general. NLO is a field that can actively employ plasmonic materials, from the few reports reviewed here on the possibilities and unaddressed features of this research area.

The phenomenon of oxidation has in this review been revealed to be beneficial as pertinent to the application. The prevention (through passivation), control (through passivation/utilizing designs that self-limit oxidation), and reversal (through chemicals/fabrication processes) can go a long way in opening up possible uses of Cu especially in terms of improving their durability and even the plasmonic virtues such as the intensity and reach of the near and far fields [480]. Intense investigations of multiple configurations and morphologies of Cu as discussed in the respective sections illuminate the road ahead in terms of realizing CuNPs with a short shackle for oxygen.

Copper compounds have immense potential and have already been explored for example for unconventional applications such as plasmonic batteries [481, 482] and plasmonic electrical switches in the form of phase and composition-controlled chalcogenides for example. All of these emerging fields of PC applications can benefit from the UV, visible, and near infrared photon harvesting from the ambient solar spectrum, catalyzing and controlling the electrochemistry of these energy storage devices [483]. Such unconventional applications indeed seem to be restricted only by contemplations of appropriate design/methods of study. Cu alloys are another class of highly nascent materials, wherein only preliminary reports of the observations of differing plasmonic properties with regard to different material parameters have been done. Investigations of the electronic structure of these alloys are hardly if ever dealt with, which is crucial to gain insights on the mechanisms of plasmonic manifestation. This aspect need not be complex, as simpler tools can reveal elemental presence post which simulation tools can take over the electronic structure calculations in many cases. Considering the inaccessibility of sophisticated imaging technologies for looking at the atomic level structures of Cu-based plasmonic alloys and compounds, computational tools need to be sought. These tools have their own shortcomings from the aspect of having available accurate material constants for novel material combinations, although approximations can help considerably. It is admitted, however, that multiple metal combinations have been explored as alloying elements with Cu already.

A general observation is that reports of plasmon-enhanced properties cannot be compared in most cases and yet they are, with the resulting materials being touted as better performers. A prevalent example of this is in PT energy conversion, wherein apparent differences in irradiation powers are still compared with another published research and reported to be better. This is intuitively a roadblock to achieve materials that are benchmarked globally to be evolving candidates for such applications. A similar application that suffers from the same drawback is theranostics, wherein the parameters of irradiation intensity for therapy vary and yet are compared against one another. Consistent comparisons or parameters hence need to be established, which when quantified for the different materials for a particular purpose nullify differences, thereby serving as a universal metric for establishing the best materials. One such parameter is the refractive index sensitivity, which carries the same definition across studies. Another parameter is the Keldysh constant, which allows confirmation of the plasmonic contributions from a material. Such variables are invariably needed especially in the fields of theranostic and PT energy storage, as the comparability between investigations can only be correct if done through them. With the promise held by Cu in plasmonics and its increasing participation as a frontrunner in this field, it stands that multiple avenues of research and commercialization will see Cu as a very good option in the realms of the ultra-violet to the near infrared, with even extensions to the microwave regime [484].